Set of integrated grid capacitors and method of manufacturing the same

0001The invention relates to an integrated grid capacitor arrangement, which contains at least one circuit-technically effective main part.

0002From an integrated arrangement, individual components can not be separated mechanically without destroying the components. Among others, layer deposition methods and layer patterning methods are used as manufacturing techniques for integrated devices.

0003A capacitor comprises two opposing electrodes, between which a dielectric is arranged. Examples of integrated capacitors are:<ul id="ul0001" list-style="dash" compact="compact"><li>so-called MIM capacitors (metal insulator metal),</li><li>stacked capacitors, which are also referred to as sand (wich) capacitors, or</li><li>Grid capacitors, also referred to as grid capacitors.</li></ul>

0004A capacitor is effective in terms of circuitry if it is not only parasitic, ie actually undesirable, but is also required for the functioning of the circuit arrangement. For example, capacitively effective capacitors serve as:<ul id="ul0002" list-style="dash" compact="compact"><li>Block or backup capacitor,</li><li>Part of a resonant circuit,</li><li>Charging capacitor, or</li><li>for storing digital information.</li></ul>

0005In the production of so-called BEOL capacities (Back End Of Line) or Far-BEOL capacity in integrated form, there are considerable variations in the capacitance values. The scattering is caused by geometry deviations due to process inhomogeneities. The scattering occurs within a semiconductor wafer or a wafer, within a production lot as well as between different production lots. If the capacitance value of a capacitor lies outside the specified specification limits, the result is a so-called performance loss or even a loss in yield of the corresponding integrated circuit.

0006From the <patcit id="pcit0001" dnum="EP1111694A2"><text>EP 1 111 694 A2</text></patcit> is a component with thin-film circuit and trimmable capacitor known. A bottom electrode of the capacitor has a structured surface with slots defining finger-shaped sections. The finger-shaped sections can be separated by means of a laser. A cover electrode may have finger-shaped sections or be made without finger-shaped sections.

0007From the <patcit id="pcit0002" dnum="US3766308A"><text>US 3,766,308</text></patcit> a capacitor arrangement is known which contains a main capacitor and a plurality of switchable auxiliary capacitors. From the<patcit id="pcit0003" dnum="US5659182A"><text>US 5,659,182</text></patcit> For example, a fuse having three terminals is known.

0008It is an object of the invention to provide a simple to produce set of grid capacitor arrangements whose capacity is as close as possible to a predetermined nominal capacity.

0009This object is achieved by a set of capacitor assemblies having the features specified in claim 1 and by an associated method having the features specified in claim 4. Further developments are specified in the subclaims.

0010A set used for a better understanding of the invention contains at least two integrated capacitor arrangements, which have been produced according to the same geometric designs or layouts and which each contain a circuit-technically effective main capacitor and at least one correction capacitor. The one capacitor arrangement contains an electrically conductive connection between the correction capacitor and the main capacitor, wherein the connection has been produced after the production of the main capacitor of this capacitor arrangement. The other capacitor assembly includes an electrically insulating interruption between the same correction capacitor and the main capacitor, the interrupt being made in accordance with the geometric designs.

0011If, for example, corrections have to be made only for 20 percent of the capacitor arrangements that have been produced in order to achieve the setpoint capacitance, the complexity drops considerably in comparison with a correction possibility in which the correction can be carried out only by disconnecting capacitors. With such a correction, capacitors would have to be locally separated in 80 percent of the capacitor arrangements.

0012In a further development, the connection is produced in a simple manner by local heating, so that components in the vicinity of the connection are protected against increased heat load. In addition, by local heating in a simple way permanent links can be produced. Compounds can be made by local heating at lower temperatures compared to making interruptions by local heating. Thus, the thermal stress, in particular in the production of a plurality of links on a semiconductor wafer is low.

0013The local heating is carried out, for example, by means of a laser beam. Thus, there is a possibility of increasing or reducing the capacitance of the circuit-technically active main capacitor by means of a parallel or series connection with the correction capacitors. This possibility of correction makes it possible to produce integrated capacitor arrangements with predetermined capacitance values ​​in a simple way.

0014In one development, the connection contains two interconnected conductor track sections, between which only one dielectric is arranged. The materials of the interconnect sections and of the dielectric are chosen so that material warping of the interconnect occurs upon heating, which penetrate the dielectric. For example, the otherwise undesirable "spiking" can be used here to establish an electrical connection. In an alternative embodiment, the dielectric contains dopant atoms that change the conductivity of the dielectric when heated. In this context, it is also spoken of an activation of the doping atoms. However, there are other ways to make the connection, which is also referred to as antifuse.

0015In a next development, the circuit arrangements contain at least one further correction capacitor, which is separated from the main capacitor or has been connected to the main capacitor.

0016The further development is based on the consideration that the capacitance values ​​of integrated capacitors which are actually to be built up at the same time can spread around a desired value both downwards and upwards. Corrections can be kept to a minimum if it gives both a possibility of correction of the capacity value towards larger capacity values ​​and a possibility of correction to smaller capacity values. The further correction capacitor offers this second correction option in comparison to the above-mentioned correction capacitor with parallel connection of the capacitors.

0017In a further development, a further interruption for separating the further correction capacitor is produced by local heating. The local heating can be carried out, for example, by means of a laser beam or by means of a current surge through a track narrowing.

0018In another development, dielectrics of the capacitors are formed by a dielectric between metallization layers in which connection sections of connections to integrated semiconductor components of the integrated capacitor arrangement are located. In the region of the capacitor, a different dielectric can be used than in the remaining region between the metallization layers, for example a dielectric with a higher dielectric constant. Examples of such capacitors are stacked capacitors or grid capacitors. In other words, the electrodes of a capacitor lie in more than two metallization layers. In such capacitors, not only electrodes in the uppermost metallization layer but also electrodes in lower metallization layers during the correction are circuitry separated or circuitry added in one embodiment. The connecting elements and interruptions for the lower electrodes lie either in the lower metallization layer, so that, for example, correspondingly deep recesses for a laser beam are to be provided, or in an upper metallization layer to which connections lead from the lower metallization layer.

0019In another embodiment, dielectrics of the capacitors have a thickness that is substantially smaller than the thickness of the dielectric between metallization layers. Examples of such capacitors are MIM capacitors. In other words, at least one electrode of the capacitor is outside a metallization layer.

0020In a next development, the capacitance of a correction capacitor is less than 1/3, less than 1/10, less than 1/100 or less than 1/1000 of the capacity of a main capacitor. By this measure it is achieved that a fine trimming is possible. Thus, with correction possibilities in the per thousand range, the capacitance values ​​of two capacitor arrangements of the same integrated circuit arrangement can be coordinated very precisely with one another. This is absolutely necessary for some applications.

0021The invention relates to a set of lattice capacitors with correction transverse electrodes. Also, individual transverse electrodes of the grid capacitors offer a possibility for a correction of the capacitance in the production. The above-mentioned technical effects therefore apply in particular to the grid capacitors.

0022In the following, embodiments of the invention will be explained with reference to the accompanying drawings. Show:<ul id="ul0003" list-style="none"><li><figref idref="f0001">FIG. 1</figref> a top view of an MIM capacitor assembly,</li><li><figref idref="f0002">FIG. 2</figref> a plan view of a grid capacitor arrangement,</li><li><figref idref="f0003">FIG. 3</figref> a transverse electrode of a grid capacitor with two interruption possibilities,</li><li><figref idref="f0003">FIG. 4</figref> a transverse electrode of a lattice capacitor with a continuous interruption capability, and</li><li><figref idref="f0004">FIG. 5</figref> Method steps for correcting the capacitance of an integrated capacitor.</li></ul>

0023<figref idref="f0001">FIG. 1</figref> 11 shows a plan view of an MIM capacitor arrangement 10, which includes a main capacitor 12 connected to an electronic circuit and a plurality of separable capacitors located to the right of the main capacitor 12, of which only a separable capacitor 14 is shown. The dielectric is in<figref idref="f0001">FIG. 1</figref> not shown for reasons of clarity.

0024In the capacitor arrangement 10 are left of the main capacitor 12 a plurality of switchable capacitors, of which in <figref idref="f0001">FIG. 1</figref> a switchable capacitor 16 is shown. The capacitors 12 to 16 are constructed the same except for their longitudinal dimensions. The capacitors 12, 14 and 16 each contain a substrate near the bottom electrode 18, 20 and 22 and a substrate remote electrode 24, 26 and 28. The bottom electrodes 18, 20 and 22 are longer than the respectively associated substrate remote electrode 24, 26 and 28 and project in the longitudinal direction on both sides of the substrate remote electrode 24, 26 and 28 addition, so that at the substrate near bottom electrodes 18, 20 and 22 connecting areas for vertically extending contacts 30 arise.

0025The main capacitor 12 and the separable capacitors 14 are electrically connected in parallel by interconnects 32 between the bottom electrodes 18, 20 and by interconnects 34 between the substrate-remote electrodes 24, 26. Geometrically, however, the separable capacitors 14 are arranged in a row one behind the other. The interconnects 32 and 34 are in an upper metallization. To each interconnect 32 and 34 performs a recess 36 and 38 in an insulating material, not shown, which covers the interconnects 32 and 34. By means of the cutouts 36 and 38, when the capacitance of the main capacitor 12 is corrected, regions 40 and 42 of a conductive path 32 or 34 can be vaporized with the aid of a laser beam, so that an interruption occurs.

0026On the other hand, the switchable capacitors 16 by means of interconnects 52 between the bottom electrodes 18, 22 and interconnects 54 between the substrate remote electrodes 24, 28 to the main capacitor 12 are electrically connected in parallel. Geometrically, however, the switchable capacitors 16 are arranged in a row. The interconnects 52 and 54 are in an upper metallization.

0027Recesses 56 and 58 lead through an insulating material, not shown, up to connecting regions 60 and 62 of the interconnects 52 and 54, respectively. The interconnecting regions 60, 62 form so-called antifuses, ie a connection possibility which, upon the impingement of a laser beam, forms a permanent electrically conductive connection between the interconnects Sections of the conductive track 52 and a conductive track 54 forms.

0028The main capacitor 12 has, in the longitudinal direction of the capacitor arrangement 10, a length La which is greater than the lengths Lb of the separable capacitors 14 or the lengths Lc of the switchable capacitors 16. Thus, a capacitance C (0) of the main capacitor 12 is greater than a capacitance Cm (I) of the capacitor 14. Here, I is a natural number for designating the last separable capacitor 14th The separable capacitors, not shown, have capacitances Cm (1) to Cm (I-1) which in the illustrated embodiment are equal to the capacitance Cm (I).

0029The capacitance C (0) of the main capacitor 12 is also greater than the capacitance Cp (1) of the switchable capacitor 16. The in <figref idref="f0001">FIG. 1</figref> Switchable capacitors not shown have capacitances Cp (2) to Cp (N) which are equal to the capacitance Cp (1) and Cm (I), respectively. N is a natural number to denote the last switchable capacitor.

0030The correction of the capacitance of the capacitor arrangement 10 will be described below with reference to FIG <figref idref="f0004">FIG. 5</figref> explained in more detail.

0031In another embodiment, there are only one separable capacitor 14 and a switchable capacitor 16 to the right and left of the main capacitor 12. In a next embodiment, there are only one or more switchable capacitors 16 or only one or more turn-off capacitors in addition to the main capacitor 12 14. In a next embodiment, the capacitances Cm (1) to Cm (I) of the capacitor array 10 are different from each other. The capacitances Cp (1) to Cp (N) of the capacitor arrangement 10 can also be formed differently from one another.

0032<figref idref="f0002">FIG. 2</figref> FIG. 12 shows a plan view of a grid capacitor assembly 110 including a main capacitor 112 connected to a circuit and a plurality of separable capacitors, of which FIG <figref idref="f0002">FIG. 2</figref> a capacitor 114 is shown. Furthermore, the grid capacitor arrangement 110 contains a plurality of switchable capacitors, of which in<figref idref="f0002">FIG. 2</figref> a capacitor 116 is shown. In<figref idref="f0002">FIG. 2</figref> Only the upper electrode of the main capacitor 112, or of the capacitor 114 or of the capacitor 116 are shown. In the underlying metallization layers, there are other electrodes that have the same course as the upper electrodes. In a first exemplary embodiment, the electrodes of a capacitor 112, 114 and 116 located in the different metallization layers are connected to one another in each case by at least one vertical contact. Trains between capacitors 112, 114 and 116 are only in the upper metallization.

0033The main capacitor 112 includes two comb-shaped electrodes in the upper metallization layer whose teeth engage with each other. From a longitudinal electrode 118 branch off, for example, four transverse electrodes 120 to 126, which are lined up in a uniform grid. On one of the longitudinal electrodes 118 opposite longitudinal electrode 128 are transverse to the longitudinal electrode 128 as many transverse electrodes 130 to 136 as on the longitudinal electrode 118, ie four transverse electrodes 130 to 136, which extend into the spaces between the transverse electrodes 120 to 126. Thus, a meander-shaped gap is formed between the transverse electrodes 120 to 126 and 130 to 136, which is filled by a dielectric. The length of the meander of the main capacitor 112 is again La.

0034The separable capacitors or the switchable capacitors are constructed like the main capacitor 112, but contain shorter longitudinal electrodes 140, 142, 144 and 146, of which, for example, only two transverse electrodes 150 to 160 branch off.

0035Between the longitudinal electrode 118 and the longitudinal electrode 142 of the capacitor 114 is a region 162, to which a recess 164 leads from the surface of the integrated circuit. Between the longitudinal electrode 128 and the longitudinal electrode 140 of the capacitor 114 is a region 166, to which a further recess 168 leads. Through the recesses 164 and 168 through an interruption between the longitudinal electrodes 118 and 142 or by means of a laser beam the longitudinal electrodes 128 and 140 are generated in the region 162 and 166, respectively. As a result, all turn-off capacitors 114 can be separated from the main capacitor 112. Between the other turn-off capacitors are more areas 170 and 172, which lead to recesses 174 and 176, so that can be separated at other locations.

0036Between the longitudinal electrode 118 and the longitudinal electrode 146 of the correction capacitor 116 is a linkage region 180. Between the longitudinal electrode 128 and the longitudinal electrode 144 of the capacitor 116 is a linkage region 182. To the linking region 140 and to the linking region 182 leads a recess 184 or 186. The recesses 184 and 186 allow the linking regions 180 and 182 to be locally heated with the aid of a laser beam. During local heating, a connection is created between the longitudinal electrode 118 and the longitudinal electrode 146 or between the longitudinal electrode 128 and the longitudinal electrode 144. By means of further linking regions 190, 192, to which recesses 194 and 196 lead, electrically conductive connections can be made to further switchable capacitors 116.

0037The meander of the main capacitor 112 has a length La larger than a length Lb of a meander of the separable capacitor 114 and a length Lc of a meander of the switchable capacitor 116, respectively. Thus, a capacitance C (0) of the main capacitor 112 is larger than a capacitance Cm (I) of the separable capacitor 114. Further capacitances Cm (1) to Cm (I-1) of further separable capacitors are equal to the capacitance Cm (I). A capacitance Cp (1) of the switchable capacitor 116 is equal to the capacitance Cm (1). Capacities Cp (2) to Cp (N) of the other separable capacitors 116 are equal to the capacitance Cp (1).

0038The correction of the capacitance of the circuit arrangement 110 will be described below with reference to FIG <figref idref="f0004">FIG. 5</figref> explained in more detail.

0039In another embodiment, there are recesses which lead to different metallization layers, for example to connection or disconnection possibilities, which are arranged offset to above or below connection or disconnection possibilities. In an alternative embodiment, link areas or Interrupt regions are located in an upper metallization layer, although they also affect electrodes disposed in lower metallization layers. Furthermore, the capacitances Cm (1) to Cm (I) and Cp (1) to Cp (N) of the capacitor array 110 may have different capacities from each other.

0040<figref idref="f0003">FIG. 3</figref> shows a transverse electrode 200 of a lattice capacitor. The transverse electrode 200 has a region 202 at approximately one third of its length and a region 204 at approximately two thirds of its length. A region 206 leads to the region 202. A recess 208 leads to the region 204. In the correction of the capacitance of the lattice capacitor, to the If the cross-electrode 200 is heard, either an interruption in the region 202 or an interruption in the region 204 is generated. If the interruption is generated in the area 202, then only about one third of the transverse electrode 200 is effective in terms of circuit technology. On the other hand, if an interruption is generated in region 204, then about two thirds of transverse electrode 200 are still effective in terms of circuit technology. By selecting a region 202 or 204, corrections of the capacitance of the lattice capacitor in the per thousand range can be carried out.

0041The length of the transverse electrode 200 is for example ten micrometers. The width is, for example, at 0.5 microns, so that a cutting with a laser beam is easily possible.

0042<figref idref="f0003">FIG. 4</figref> shows a transverse electrode 220, to which a recess 222 leads. The recess 222 extends approximately over the entire length of the transverse electrode 220. This makes it possible to interrupt the transverse electrode 220 at any point. In other words, the break points can be set continuously along the longitudinal axis of the transverse electrode 220.

0043In other embodiments of transverse electrodes 200 and 220, link areas are used in place of or in combination with the interruption areas. The interruption regions or the connection regions are arranged either on a transverse electrode of a lattice capacitor or on a plurality of transverse electrodes of the lattice capacitor.

0044<figref idref="f0004">FIG. 5</figref> shows method steps for correcting the capacitance of an integrated capacitor arrangement, for example <figref idref="f0001">FIG. 1</figref>, <figref idref="f0002">2</figref>, <figref idref="f0003">3 or 4</figref>, In advance of the method, the scatter of the capacitance of the integrated circuit arrangement around a nominal capacitance is determined in a design and simulation stage, for example empirically or based on simulation runs, see method step 300. Depending on the capacity variations, correction options are provided, see method step 302. The correction options are, for example, separable capacitors, switchable capacitors, separable capacitor areas and / or switchable capacitor areas. The correction options are predetermined in method step 302, taking into account the expected capacity dispersion, so that with respect to the entire production, as few interruptions and connections as possible must be produced by heating with the laser beam.

0045The actual correction process begins in a method step 304, which is followed by the processing of a wafer, see method step 306. For example, transistors are produced in a semiconductor material of the wafer. Thereafter, metallization layers are applied, wherein capacitors are also produced.

0046In a method step 308, the actual capacitance of an integrated capacitor arrangement is detected by means of a measurement, ie in particular the capacitance of the main capacitor 12, 112 together with the capacitances of the separable capacitors 14, 114.

0047In a following method step 310, the actual capacity is compared with the desired capacity. If the actual capacity is smaller or larger than the setpoint capacity, in particular smaller or larger than a predefined tolerance range, a method step 312 follows immediately after method step 310. In method step 312, it is checked whether the actual capacity is greater than the setpoint capacity. Capacity is. If this is the case, then, after method step 312, a method step 314 follows, in which interruptions in the integrated capacitor arrangement are generated with the aid of a laser beam, detachable capacitors 14, 114 being separated from the main capacitor 12, 112. The capacitance of the capacitor arrangement becomes smaller. Separation of an electrode section from a transverse electrode is alternatively performed.

0048If, on the other hand, it is determined in method step 312 that the actual capacity is smaller than the setpoint capacity, method step 312 is immediately followed by a method step 316. In method step 316, link areas are heated by means of a laser beam. As a result, switchable capacitors 16, 116 are connected to the main capacitor 12, 112 or to a main area. Thus, the capacitance of the capacitor arrangement increases in the direction of the desired capacity. An addition of electrode sections of a transverse electrode is alternatively carried out.

0049If, however, it is determined in method step 310 that actual capacity and setpoint capacity coincide, then a method step 318 follows immediately. Method step 318 is also carried out after method step 314 or method step 316. In process step 318, the wafer is further processed. Among other things, a passivation layer is applied, which closes the recesses for the laser beam.

0050In a further method step 320, the circuits arranged on the wafer are singulated and cast into housings. The method is ended in a method step 322.

0051In another embodiment, the correction is performed after the separation of the circuits. For example, surges are used to heat the interruption areas or the linking areas. The queries in method steps 310 and 312 can also be formulated differently.

0052The procedure given gives the following advantages:<ul id="ul0004" list-style="dash" compact="compact"><li>the variations in the capacitances caused by geometry or process fluctuations can be corrected subsequently in a simple manner. Thus, the yield or performance can be increased.</li><li>The possibility of subsequent correction also allows for the respective integrated circuit customizing the capacity. This allows a specially adapted to other circuit elements capacity adjustment can be made, for example, for optimal operating point setting. This is particularly advantageous if the other circuit elements are no longer correctable.</li></ul>

0053By separating or adding regions of the capacitor with the aid of so-called fuses or antifuses, the influence of process and geometry fluctuations on the parameters of the main capacitor, in particular on the capacitance and on the resistance (and thus on the RC constant) can be determined. , correct later.

0054When correcting in steps 310-316, the following formula can be used:<maths id="math0001"><math display="block"><mi>Ckorr</mi><mo mathvariant="normal">=</mo><mi mathvariant="normal">C</mi><mfenced><mn mathvariant="normal">0</mn></mfenced><mo mathvariant="normal">-</mo><mi>total</mi><mspace width="1em" /><mfenced><mi mathvariant="normal">i</mi><mo mathvariant="normal">=</mo><mn mathvariant="normal">1</mn><mspace width="1em" /><mi>to I</mi><mo>⁢</mo><mn mathvariant="normal">1</mn><mspace width="1em" /><mi mathvariant="normal">about Cm</mi><mspace width="1em" /><mfenced><mi mathvariant="normal">i</mi></mfenced></mfenced><mo mathvariant="normal">+</mo><mi>total</mi><mspace width="1em" /><mfenced><mi mathvariant="normal">n</mi><mo mathvariant="normal">=</mo><mn mathvariant="normal">1</mn><mspace width="1em" /><mi>to N</mi><mo>⁢</mo><mn mathvariant="normal">1</mn><mspace width="1em" /><mi mathvariant="normal">over Cp</mi><mspace width="1em" /><mfenced><mi mathvariant="normal">n</mi></mfenced></mfenced><mo mathvariant="normal">,</mo></math><img file="EP1587145B1_D0001.tif" /></maths> the sizes used have been explained above except for the sizes I1 and N1. The index i at summation over the capacitance Cm must take into account all the capacitors cut off, where I1 denotes the last disconnected capacitor. The index in the summation over the capacitances Cp must take into account all connected capacitances, where N1 designates the last connected capacitor. It should also be noted that either capacitors are switched on or disconnected.

0055If the capacitances Cm and Cp are the same, instead of the sum, it is possible to multiply by a factor which indicates the number of separated or switched capacitors.

0056When determining the capacity dispersion, you can use the following relationships that apply to a disk capacitor:<maths id="math0002"><math display="block"><mi mathvariant="normal">C</mi><mo mathvariant="normal">=</mo><mi mathvariant="normal">ε</mi><mo>⁢</mo><mn mathvariant="normal">0</mn><mo mathvariant="normal">⋅</mo><mi mathvariant="normal">.epsilon..sub.R</mi><mo mathvariant="normal">⋅</mo><mi>aeff</mi><mo mathvariant="normal">/</mo><mi>deff</mi><mo mathvariant="normal">,</mo></math><img file="EP1587145B1_D0002.tif" /></maths> where ε0, εr represent the corresponding dielectric constant, Aeff the effective electrode area and Deff the effective electrode spacing. A fluctuation in the area ΔAeff or the distance ΔDeff then causes a change in the capacity of:<maths id="math0003"><math display="block"><mi mathvariant="normal">.DELTA.C</mi><mo>=</mo><mi mathvariant="normal">ΔAeff</mi><mo>/</mo><mi>Aeff or</mi><mn>.</mn><mspace width="1em" /><mi>.DELTA.C</mi><mo>=</mo><mi mathvariant="normal">ΔDeff</mi><mo>/</mo><mi>deff</mi><mn>.</mn></math><img file="EP1587145B1_D0003.tif" /></maths>

0057In a first approximation, the effective electrode area corresponds to:<ol id="ol0001" compact="compact"><li>a) the MIM capacitors: the overlapping surface of the electrodes,</li><li>b) the sandwich capacitors: the area of ​​the metal electrodes,</li><li>c) the lattice capacitors: the side surface of the tracks, which results from the length L and the thickness T.</li></ol>

0058The effective electrode distance corresponds to:<ol id="ol0002" compact="compact"><li>a) the MIM capacitors: the thickness of the dielectric, which differs from the dielectric between the metallization layers,</li><li>b) the sandwich capacitors: the thickness of the intermetal dielectric,</li><li>c) the lattice capacitors: the thickness of the intrametal dielectric, ie the so-called Spacings.</li></ol>

0059Another influencing factor is, for example, the formation of corners in lattice capacitors. These influencing factors can at least be measured and then taken into account.

0060The fluctuations of the capacities can have the following process-related causes:<ol id="ol0003" compact="compact"><li>a) MIM capacitors: variations in the thickness of the MIM dielectric, eg due to inhomogeneous deposition rates over the wafer or different roughness of the lower electrode.</li><li>b) Sandwich capacitors: variations in the thickness of the intermetal dielectric, eg thickness variations due to polishing inhomogeneities or variations in the etching depth over the wafer.</li><li>c) Lattice capacitors: variations in the electrode spacing due to variation in the thickness of the intrametallic dielectric caused, for example, by lithographic fluctuations, RIE structuring (Reactive Ion Etching) when using aluminum or by trench etching when using copper. Further causes are fluctuations of the electrode surface by variation of the conductor track thickness or due to chemical mechanical polishing (CMP) fluctuations, so-called dishing, non-rectangular trench profile when using copper or through inhomogeneous deposition rates.</li></ol>

0061In the case of MIM capacitors and sandwich capacitors, the capacitance is corrected by directly disconnecting or connecting M surface segments of the electrode:<maths id="math0004"><math display="block"><mi mathvariant="normal">A</mi><mo mathvariant="normal">=</mo><mi mathvariant="normal">A</mi><mo>⁢</mo><mn mathvariant="normal">0</mn><mo mathvariant="normal">±</mo><mo mathvariant="normal">+</mo><mi>total</mi><mspace width="1em" /><mfenced><mi mathvariant="normal">i</mi><mo mathvariant="normal">=</mo><mn mathvariant="normal">1</mn><mspace width="1em" /><mi mathvariant="normal">to M over Ai</mi></mfenced><mo mathvariant="normal">,</mo></math><img file="EP1587145B1_D0004.tif" /></maths> where A0 is the electrode surface which can no longer be corrected, M is a natural number and Ai is the separable or switchable discrete area elements. For the separable surface elements, the minus sign applies. For the switchable surface elements, the plus sign applies.

0062In the case of the grid capacitors, the electrode area A is set over the length L of the interconnect: A = L. T, where T is the mean interconnect thickness within the capacitor structure. By separating or connecting M track segments, eg n discrete segments Li of the electrode is corrected:<maths id="math0005"><math display="block"><mi mathvariant="normal">L</mi><mo mathvariant="normal">=</mo><mi mathvariant="normal">L</mi><mo>⁢</mo><mn mathvariant="normal">0</mn><mo mathvariant="normal">±</mo><mo mathvariant="normal">+</mo><mi>total</mi><mspace width="1em" /><mfenced><mi mathvariant="normal">i</mi><mo mathvariant="normal">=</mo><mn mathvariant="normal">1</mn><mspace width="1em" /><mi mathvariant="normal">to M over Li</mi></mfenced><mo mathvariant="normal">,</mo></math><img file="EP1587145B1_D0005.tif" /></maths> where L0 denote the basic electrode length which can no longer be corrected, M a natural number and Li the separable discrete track segments. For the separable track segment, the minus sign applies. The plus sign applies to the switchable track segments.