Patent Application
20250137140
The present invention relates to the field of semiconductors, in particular three-dimensional devices such as dynamic random access memory (DRAM), integrated circuits, and three-dimensional vertical NAND storage memories (3D-NAND).
Dynamic random access memories consist of a transistor cell and a capacitor that are based on metal-oxide-semiconductor (MOS) technology. Capacitors are usually MIM (Metal-Insulator-Metal) capacitors composed of two U-shaped electrodes, an upper electrode and a lower electrode, separated by a dielectric material. The electrodes are based on metal oxides (i.e., RuO2 and SrRuO3, SnO2 and MoO2) or metal (i.e., Pt, TiN, TaN, Ru—Pt, SnO2 doped Ta). The dielectric can be, for example, Al2O3, TiO2, ZrO2, HfO2, SrTiO3 (STO) or BaSrTiO3 (BST) or ZrO2 stabilized by Y. The electrodes are manufactured by atomic layer deposition (ALD) which is the only approach that has been able to achieve perfect compliance on a complex shape structure, such as the U-shape. The ALD method is also the only one that can obtain thin upper electrodes. However, these methods reach limits at technological nodes of 100 nm. In addition, they must be optimized to increase capacitance density, to perfectly fill cavities and to limit current leakage, such as by reducing the thickness of the metal electrodes without reducing conductivity.
3D-NAND memory devices consist of a semiconductor substrate on which is arranged a stack of alternating layers of conductive material, called “word lines,” which are metallic in nature, and layers of insulating material made of an inorganic dielectric. On the sides of the device, the stack may be etched in a “staircase” pattern, with the length of the layers and the number of layers decreasing from one stage to the next in the upward direction.
In current fabrication processes, word lines are created by simultaneously filling several rows of long, thin horizontal cavities separated by insulating layers with a conductive material, the most commonly used conductive material being tungsten. Tungsten is usually deposited in two steps: an atomic layer deposition (ALD) step to create a thin cling layer, followed by a chemical vapor deposition (PECVD) step to completely fill the cavities. However, the further down the stack one goes, the more difficult it becomes to fill them completely. This problem is amplified when the word lines to be fabricated are thinner and the cavity opening is smaller. As a result, material voids can be formed in the tungsten deposits, leading to conductivity losses and memory malfunctions. Thus, this filling technology has shown its limits, especially for stacks comprising 96 or 128 word lines.
Moreover, prior art vapor deposition processes tend to deform the silicon coupon and cause it to bow after annealing at the end of the process. To avoid this deformation, it has been proposed to deposit a layer on the back side of the wafer (the front side being covered by the stacks) by performing a plasma-enhanced chemical vapor deposition (PECVD) process followed by wet etching. However, this method has the disadvantage of adding at least two steps in the fabrication process of the memory device and requires controlling the treatment of the back side without damaging the front side in the following steps.
Consequently, there is a need to provide an improved 3D NAND memory, in particular to fabricate a 3D NAND memory in which the word lines have a higher conductivity with the same number of stages, the increase in conductivity being achieved by adjusting at least one of the following two parameters: (1) the reduction of the material voids in the lines and/or (2) the optimization of the volume of conductor available to create the lines.
Previously, when the word lines were tungsten based, it had been necessary to interpose between the metal lines and the blocking dielectric (generally SiO2 optionally associated with alumina Al2O3), a thin layer of a “barrier” material which prevents the migration of the elements contained in the metal towards the dielectric. The layer of barrier material reduces the space available for the metal, the height of the spaces to be filled being imposed. An example of material combinations that have been interposed between the polysilicon channel and the tungsten lines is: polySi/SiO/SiN/SiO2/Al2O3/TiN/W. The multiplication of the number of successive layers deposited to create the word lines considerably limits the volume of tungsten conductor. In addition, the deposition of additional materials adds steps to the fabrication of the memory device.
In semiconductor devices such as 3D-NAND memories, it was suggested in WO2021/219744 to replace tungsten word lines with an alloy of nickel and boron. However, while this technology filling technology has demonstrated an improvement over the use of tungsten, further modifications are needed to create very thin conductive layers.
There remains the need for a metallization process that can be used in the manufacture of semiconductor devices, which allows the metallization of a mineral oxide substrate with metal layers that remain conductive even at very thin thickness, including thicknesses of between about 1 and about 25 nm.
Finally, in three-dimensional integrated circuits, the metal interconnects connecting the electronic components contain copper, separated from the semiconductor substrate by layers of barrier materials such as titanium nitride and tantalum nitride. It would be desirable to replace these materials to improve the performance of the devices.
It is an object of the present invention to provide an improved metallization process that can be used in the manufacture of semiconductor devices.
It is another object of the present invention to provide a metallization process that can be used for metallization of a mineral oxide substrate.
It is another object of the present invention to provide a metallization process that allows for the metallization of a mineral oxide substrate with metal layers that remain conductive at very thin thicknesses.
To that end, in one embodiment, the present invention relates generally to a method for metallizing a mineral oxide substrate with a nickel-boron or cobalt-boron alloy, said metallization process comprising:
In one embodiment, the present invention relates generally to a method of electroless deposition of an alloy of nickel and boron that includes a step of activation of a surface of a mineral oxide substrate with a noble metal. This activation step comprises:
Once the surface of the mineral oxide substrate has been activated by the first activation step followed by the second activation step, an alloy of nickel and boron can be deposited on the activated surface of the mineral oxide substrate by electroless deposition.
As used herein, “a,” “an,” and “the” refer to both singular and plural referents unless the context clearly dictates otherwise.
As used herein, the term “about” refers to a measurable value such as a parameter, an amount, a temporal duration, and the like and is meant to include variations of +/−15% or less, preferably variations of +/−10% or less, more preferably variations of +/−5% or less, even more preferably variations of +/−1% or less, and still more preferably variations of +/−0.1% or less of and from the particularly recited value, in so far as such variations are appropriate to perform in the invention described herein. Furthermore, it is also to be understood that the value to which the modifier “about” refers is itself specifically disclosed herein.
As used herein, spatially relative terms, such as “beneath,” “below,” “lower,” “above,” “upper” and the like, are used for ease of description to describe one element or feature's relationship to another element(s) or feature(s) as illustrated in the figures. It is further understood that the terms “front” and “back” are not intended to be limiting and are intended to be interchangeable where appropriate.
As used herein, the terms “comprises” and/or “comprising,” specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
As used herein, the term “substantially free” or “essentially free” if not otherwise defined herein for a particular element or compound means that a given element or compound is not detectable by ordinary analytical means that are well known to those skilled in the art of metal plating for bath analysis. Such methods typically include atomic absorption spectrometry, titration, UV-Vis analysis, secondary ion mass spectrometry, and other commonly available analytically techniques.
All amounts are percent by weight unless otherwise noted. All numerical ranges are inclusive and combinable in any order except where it is logical that such numerical ranges are constrained to add up to 100%.
As used herein, the term “alloy” means a solid solution in which the elements are evenly distributed.
In one embodiment, the alloy is a nickel alloy comprising boron.
In another embodiment, the alloy is a cobalt alloy comprising boron.
In one embodiment, the nickel-boron alloy or cobalt-boron alloy may also include at least one of phosphorus and tungsten.
In addition to nickel (or cobalt) and boron, the alloy may also contain impurities from the ingredients present in the electroless aqueous solution.
The invention also relates to a three-dimensional semiconductor device obtainable by implementing the method described above.
The method of the invention is applied in particular in the realization of the filling of cavities which have been previously formed in a semiconductor substrate, and whose dimension at their opening is less than 5 microns, or less than 3 microns or less than 1 micron or less than 500 nm or less than 100 nm or less than 50 nm or less than 25 nm or less than 10 nm.
As described herein, the method of the present invention makes it possible to create metal deposits on substrates with complex topographies in terms of relief and shape, and to create metal layers on the walls of cavities of small size at their opening and of great depth. The method of the invention makes it possible in particular to overcome the materials used in the prior art as barriers to the diffusion of copper or other metals.
The process of the invention makes it possible to deposit a metal layer directly on an inorganic dielectric material, which material may be a mineral oxide substrate. The mineral oxide substrate may be selected from silicon dioxide (SiO2), alumina (Al2O3), hafnium oxide, zirconium oxide and their silicates, by way of example and not limitation.
The process of the invention also makes it possible to not require the deposition of a barrier layer, generally of titanium nitride or tantalum nitride, more commonly of titanium nitride. The absence of a barrier material provides the advantages of eliminating a step in the process without reducing the conductivity of the metal lines. The inventors of the present invention have discovered that the nickel-boron (or cobalt-boron) alloy does not diffuse into the dielectric. The nickel-boron (or cobalt-boron) alloy used as a conductor in place of tungsten not only exhibits barrier properties, but also provides higher conductivity than tungsten.
The alloy deposition step can advantageously be performed by subjecting the mineral oxide surface to ultrasound. The mineral oxide with which the electroless solution is brought into contact may be, for example, SiO2 or Al2O3.
The nickel or cobalt alloy deposit may be formed at various stages of the 3D NAND memory fabrication process. For example, the alloy may be deposited to form a word line, to form a contact between a polysilicon channel and a bit line, or to form the barrier layer of a bit line.
In one embodiment, the process of the invention is used to metallize a mineral oxide substrate with an alloy of nickel-boron or cobalt-boron to in the process of fabricating a 3D NAND device.
The present application also describes a process for metallizing a semiconductor substrate, said substrate comprising at least one horizontal cavity opening onto a vertical cavity, the horizontal cavity having a smaller opening size than the vertical cavity, said cavities defining a surface comprising at least one area of an inorganic oxide, said metallization process comprising a step of selectively activating the inorganic oxide surface, followed by a step of depositing a metal on the activated inorganic oxide surface by bringing said surface into contact, in the absence of polarization, with an electroless solution comprising metal ions, preferably nickel or cobalt ions, in combination with a suitable reducing agent of the metal ions.
In one embodiment, the present invention relates generally to a method of metallizing one or more surfaces of a mineral oxide substrate with an alloy of nickel and boron by electroless deposition. Prior to electroless deposition, the one or more surfaces of the mineral oxide substrate are activated with a noble metal.
This activation step comprises:
The mineral oxide substrate may be selected from silicon dioxide (SiO2), alumina (Al2O3), hafnium oxide, zirconium oxide and their silicates.
In the first activation stage, the one or more surfaces of the mineral oxide substrate are contacted with a first activation solution. This first activation solution that includes a noble metal complex and an organo-silane binder in combination with one or more solvents.
In one embodiment, the noble metal may be selected from the group consisting of ruthenium, rhodium, osmium, iridium, palladium, platinum, gold and/or silver complexed with a complexing agent to form the noble metal complex. In one embodiment, the noble metal comprises palladium and the palladium is complexed with a suitable complexing agent to form a palladium complex.
In one embodiment, the palladium complex may be selected from (NH4)2(PdCl4); Pd(NH3)4 and complexes of formula (I):
In one embodiment, the palladium complex is a complex having formula (I) in which R1 and R2 represent H and X represents Cl.
In one embodiment, the first activation solution comprises the noble metal complex or complexes in a concentration of 10−6 M to 10−2 M, preferably from 10−5 M to 10−3 M, preferably from 5·10−5 M to 5·10−4 M.
According to a particular feature of the invention, the organo-silane binder used in the first activation solution may have the general formula (II):
{X-(L)}3−nSi(OR)n (II)
wherein:
The organo-silane binder may also have the general formula (III):
(OR)3Si-(L)-Si(OR)3 (III)
wherein:
L represents a spacer arm selected from the group consisting of CH2CH2CH2NHCH2CH2NHCH2CH2 and CH2CH2CH2—S—S—CH2 CH2CH2; and
Compounds of formulas (II) or (III) are for example selected from the following compounds: (3-aminopropyl) triethoxysilane; (3-aminopropyl)trimethoxysilane; m-aminophenyltrimethoxysilane; p-aminophenyltrimethoxysilane; p,m-aminophenyltrimethoxysilane; 4-aminobutyltriethoxysilane; m,p (aminoethylaminomethyl) phenethyltrimethoxysilane; N-(2-aminoethyl)-3-aminopropyltriethoxysilane; N-(2-aminoethyl)-3-aminopropyltrimethoxysilane; 2-(4-pyridylethyl)triethoxysilane; bis(3-trimethoxysilylpropyl)ethylenediamine; (3-trimethoxysilylpropyl) diethylenetriamine; N-(3-trimethoxysilyethyl)ethylenediamine; N-(6-aminohexyl)aminopropyltrimethoxysilane; (3-glycidoxypropyl) trimethoxysilane; (3-glycidoxypropyl)triethoxysilane; 5,6-epoxyhexyltriethoxysilane; (3-mercaptopropyl) trimethoxysilane; (3-mercaptopropyl)triethoxysilane; bis [3-(triethoxysilyl) propyl] disulfide; 3-chloropropyltrimethoxysilane; 3-chloropropyltriethoxysilane; (p-chloromethyl)phenyltrimethoxysilane; m,p ((chloromethyl)phenyl) trimethoxysilane.
Examples of organo-silanes binders that may be used in the context of the present invention include (3-aminopropyl)-trimethoxy-silane, (3-aminopropyl)-triethoxy-silane, [3-(2-aminoethyl) aminopropyl] trimethoxy-silane and 3-[2-(2-aminoethylamino) ethylamino] propyl-trimethoxysilane.
In one embodiment, the organo-silane binder comprises at least two or even at least three amino groups including a primary amine and two secondary amines.
The concentration of the organo-silane binder is preferably between about 10−5 M and about 10−1 M, preferably between about 10−4 M and about 10−2 M, preferably between about 5·10−4 M and about 5·10−3 M.
The solvent of the first activation solution may comprise one or more solvents selected from the group consisting of N-methylpyrrolidinone (NMP), dimethyl sulfoxide (DMSO), alcohols, ethylene glycol ethers such as monoethyl-diethylene glycol (EDEG), propylene glycol ethers, dioxane and toluene.
In one embodiment, the first activation solution comprises (ethylenediamine) palladium (II) chloride and 3-[2-(2-aminoethylamino)ethylamino]propyl-trimethoxysilane in DMSO as the solvent.
The first activation stage is preferably carried out at a temperature ranging from 50° C. to 80° C., for example from 60° C. to 70° C. The first activation stage is generally carried out for a time period of about 30 seconds to about 30 minutes, more preferably about 5 minutes to about 30 minutes. The pH of the first activation solution may be within a range of 2 to 10.
Contact of the mineral oxide substrate with the first activation solution can be carried out by soaking, optionally applying ultrasound, a coupon with the mineral oxide surface in a container containing the first activation solution described above.
After contacting the mineral oxide surface with the first solution for a sufficient time, the activated one or more surfaces of the mineral oxide substrate are rinsed to remove all traces of the first activation solution.
The first activation stage leads to the formation of an activated surface. This first activation stage is then followed by a second activation stage in which the activated surface is contacted with a second aqueous activation solution having a pH ranging from 1 to 4 and comprising noble metal ions. In one embodiment, the noble metal ions are selected from the group consisting of ruthenium ions, rhodium ions, osmium ions, iridium ions, palladium ions, platinum ions, gold ions, silver ions, and combinations of the foregoing. In one embodiment, the noble metals ions comprise palladium ions, preferably Pd2+ ions. In one embodiment, the noble metal used in the first activation solution is the same as the noble metal used in the second activation solution.
The second activation solution may be prepared by dissolving a source of the noble metal ions in water and then adjusting the pH to between 1 and 4, preferably between 1.5 and 2.5. In one embodiment, the source of the noble metals ions is a compound selected from the group consisting of tetrachloropalladic acid, palladium trifluoroacetate (II), palladium pivalate (II), palladium (II) trimethylacetate and palladium acetate (II). In one embodiment, the concentration of the source of noble metal ions is in the range of about 10−8 M to about 10−1 M, more preferably about 10−6 M to about 10−2 M. If necessary, a pH adjuster may be added to the solution, including, for example a mineral acid such as sulfuric acid, nitric acid, phosphoric acid or hydrochloric acid.
The second activation stage is carried out at a temperature in the range of about 15° C. to about 40° C., more preferably about 20° C. to about 25° C. The duration of the second stage can be between 5 seconds and 60 seconds. Contact of the activated mineral oxide surface with the second aqueous solution can be carried out by soaking in the second activation solution described above.
After contacting the activated mineral oxide surface with the second aqueous solution for a sufficient time, the activated mineral oxide surface is rinsed to remove all traces of the second solution.
Without wishing to be bound by theory, the inventors of the present invention believe that the two-stage activation step described herein enables complete activator coverage of the vertical substrate surfaces enabling, at very low thicknesses, a continuous conductive layer for subsequent electroless deposition.
That is, while our U.S. Provisional Application No. 63/307,234 requires both boron and either phosphorus or tungsten in the alloy at the recited atomic percentage (at. %) to produce a deposit that is more conductive, especially at very low thicknesses, the present invention does not require the presence of tungsten or phosphorus in the alloy to achieve such a conductive deposit at extremely low thicknesses. That is, in one embodiment, the nickel-boron or cobalt-boron alloy described herein is at least essentially free of tungsten and phosphorus, meaning that the alloy does not contain tungsten or phosphorus beyond an amount present as impurities from ingredients present in the electroless plating solution. In another embodiment, the nickel-boron or cobalt-boron alloy contains small amounts of tungsten or phosphorus, such as less than 0.1 at. % or less than 0.01 at. % or less than 0.001 at. % tungsten and/or phosphorus.
Once the two stage application step has been completed, the activated mineral oxide substrate may be contacted with an electroless plating solution containing metal ions, preferably one or more of nickel ions and cobalt ions along with a reducing agent comprising boron to deposit a nickel-boron or cobalt-boron alloy on the activated surfaces of the mineral oxide substrate. In one embodiment, this may be accomplished by contacting the activated mineral oxide surface with an electroless aqueous solution comprising the nickel ions or cobalt ions and a nickel ion or cobalt ion reducing agent comprising boron.
The contact is carried out, for example, under conditions allowing the formation of a layer of nickel-boron alloy or a cobalt-boron alloy having a thickness of at least 5 c, for example a thickness ranging from 5 nm to 20 nm, or from 7 nm to 15 nm, or from 7 nm to 10 nm on surfaces of the mineral oxide substrate.
The average resistivity of this nickel-boron or cobalt-boron alloy layer, measured by a method known to those skilled in the art, such as a four-point probe at different points in the sample, can range from 20 to 300 μohm·cm, preferably from 20 to 200 μohm·cm or from 20 to 150 μohm·cm.
The resistivity of the alloy layer is preferably less than 300 μohm·cm when the thickness of the layer is less than or equal to 20 nm. The resistivity of the alloy layer is for example between 80 and 120 μohm·cm when the thickness of the layer is between 5 nm and 15 nm.
The method of the invention makes it possible to obtain a layer of nickel-boron alloy or cobalt-boron alloy which is conductive when its thickness is between 5 nm and 14 nm. The term “conductive layer” within the meaning of the present invention, refers to a layer for which a resistivity value is obtained by a measurement method known to those skilled in the art, for example by the four-point method. For comparison, a method of the prior art wherein the activation step comprises only the first step described above, does not permit layers of nickel alloy and boron to be formed which are conductive when their thickness is less than or equal to 14 nm.
In a particular embodiment, the method of the invention makes it possible to obtain a layer of nickel alloy and boron having a lower thickness than a metal layer obtained according to the teachings of WO2021/219744, with values of equal conductivities.
The reducing agent of nickel ions comprising boron is preferably used in a sufficient quantity such that boron represents between 0.1 at. % and 10 at. % in the nickel or cobalt alloy, more preferably about 0.1 at. % to about 5.0 at. %, more preferably between about 0.1 and about 3.0 at. %. Nickel (or cobalt) ions are present in sufficient quantities for nickel (or cobalt) to account for between 90% and 99.9 at. %, more preferably about 95 at. % to about 99.9 at. %, more preferably about 97.0 at. % to about 99.9 at. % in the nickel-boron or cobalt-boron alloy.
The electroless aqueous solution can be obtained by dissolving in water a nickel metal salt preferably selected from the group consisting of acetate, acetylacetonate, hexafluorophosphate, nitrate, perchlorate, sulfate or nickel tetrafluoroborate. A hydrated form of one of listed nickel salts may also be used. In one embodiment, the nickel salt is nickel sulfate hexahydrate. The nickel ions are in a concentration for example between 10−2 M and 1 M, preferably between 50 mM and 500 mM.
In the case of an electroless cobalt solution, the cobalt ions are introduced into the electroless solution as an inorganic cobalt salt such as chloride and/or sulfate or other inorganic salts or inorganic complexes such as pyrophosphates or a cobalt complex with an organic carboxylic acid salt such as acetate, citrate, lactate, succinate, propionate, and hydroxyacetate. The cobalt ions are in a concentration for example between 102 M and 1 M, preferably between 50 mM and 500 mM.
The reducing agent comprising boron may be a borane derivative such as a borane derivative selected from dimethylaminoborane, pyridine borane, morpholene borane and terbutylamine borane.
The electroless aqueous solution optionally contains at least one nickel ion stabilizing agent, which is preferably in an amount between 10−3 M and 1 M.
The electroless aqueous solution may contain a nickel ion stabilizing agent selected from the group consisting of ethylene diamine, citric acid, acetic acid, succinic acid, malonic acid, aminoacetic acid, malic acid or an alkali metal salt of these compounds. A preferred stabilizing agent in the context of the present invention is citric acid, which forms complexes with nickel ions in the solution.
The electroless aqueous solution may comprise an agent for adjusting the pH to a value between 6 and 11. In one embodiment, the pH of the aqueous solution is in the range of 8 to 10, more preferably between 9.0 and 9.5. The pH adjusting agent may be selected from the group consisting of aminoethanol, N-methyl aminoethanol and N, N-dimethyl-aminoethanol. A preferred pH adjusting agent is N-methyl aminoethanol.
The electroless aqueous solution may further contain a polyamine, preferably an aliphatic polyamine, in a concentration that may be between 1 ppm and 1000 ppm (mg/L). In a preferred embodiment, the electroless solution contains a polyethyleneimine, and preferably a polyethyleneimine of molecular weight in number greater than or equal to 500 g/mol, more preferably greater than about 600 g/mol, more preferably greater than about 700 g/mol. The solution may alternatively contain a polymer selected from derivatives of chitosan, poly (allyl-amine), poly(vinyl-amine), poly(vinyl-pyridine), poly (amino-styrene), poly (L-lysine), and acid (or protonated) forms of these polymers.
The contact of the substrate with the electroless solution can be carried out by immersing the mineral oxide substrate in the solution described above, at a temperature between 40° C. and 90° C., preferably at 60° C. to 70° C., for a period of 30 seconds to 15 minutes, more preferably about 1 to about 10 minutes, depending on the desired thickness of nickel alloy. The contact of the electroless solution is advantageously carried out for a sufficient time to obtain a nickel or cobalt alloy layer having a thickness of greater than 1 nm to less than 25 nm, more preferably greater than or equal to 4 nm to less than or equal to 10 nm.
The deposition step of the alloy metal layer can be carried out under different conditions. For example, the substrate to be coated can be rotated. A recirculation of the electroless solution can be imposed in the reactor. Contact of the substrate with the electroless solution can be carried out by wetting by spraying the solution at high pressure. Other means can be used in a complementary way by shaking for example the substrate and/or the solution with ultrasound or megasound. In all cases, the contact can be carried out under vacuum. The surface to be coated can be positioned face up or face down.
According to an advantageous embodiment, this layer can be annealed at a temperature between 200° C. and 700° C., preferably at 400° C., for a period of between 1 minute and 30 minutes, preferably about 10 minutes, under an inert or reducing atmosphere (4% hydrogen in nitrogen).
An example of an electroless aqueous solution comprises:
In one embodiment, the electroless solution consists essentially of the listed ingredients. What is meant by “consisting essentially of” is that the electroless solution is free of any ingredients that would have a detrimental effect on conductivity, including a concentration of boron, phosphorus, and/or tungsten in excess of the ranges defined herein.
In one embodiment, the electroless solution consists of the listed ingredients to provide an electroless nickel or cobalt alloy layer on a mineral oxide substrate that exhibits good conductivity at very low thicknesses.
The step of depositing the alloy metal layer may be performed under different conditions. For example, the substrate to be coated may be rotated. A recirculation of the electroless solution may be imposed in the reactor. The substrate may be brought into contact with the electroless solution by spray wetting the solution at high pressure. Other means may be used in a complementary manner, for example by agitating the substrate and/or the solution with ultrasound or ultrasound guns. In any case, the contacting may be done under vacuum.
A method for fabricating a 3D NAND memory according to the invention may include, in addition to the selective metallization process, other steps necessary to provide a memory device that is functional. Alternatively, the method of the invention may comprise steps additional to those described above, to provide a portion of the memory device only.
A first example of a 3D NAND memory obtained according to the method of the invention comprises a semiconductor substrate defining a horizontal plane, at least one semiconductor channel disposed along a vertical axis, and several word lines comprising the nickel or cobalt alloy.
The method for fabricating such a memory may include, in addition to the metallization process, at least one other step. For example, the method of the invention may comprise, prior to the steps of activation and electroless deposition of the nickel or cobalt alloy, the following steps:
These two steps are followed by a step of selective deposition of the nickel or cobalt alloy in the horizontal cavities by an electroless process described above. The selective deposition step ensures that the vertical cavity is not filled with nickel or cobalt alloy after the step is completed.
The vertical cavity may have an average opening diameter ranging from 80 nm to 150 nm, and a depth greater than 1 micron, and the horizontal cavities may have an average width along a vertical axis less than the average opening diameter of the vertical cavity.
The method of the invention provides a particularly advantageous alternative for the fabrication of 3D NAND comprising more than 90 word lines of tungsten or other physically deposited metal. A substrate used for the fabrication of such 3D NAND that is brought into contact with the electroless solution (also referred to as electrolyte) described above may comprise a layer count greater than or equal to a value selected from the group consisting of 32, 48, 64, 96, 128, 192, 256, preferably 96 or 128.
In this substrate, the vertical cavities have in particular an average diameter at the opening of less than 1 micron, for example ranging from 50 nm to 150 nm, and a depth of more than 1 micron, and the horizontal cavities have an average width along a vertical axis of less than 100 nanometers and an average depth along a horizontal axis of less than 100 nanometers.
According to a second example of the process of the invention, the nickel or cobalt alloy deposit forms at least a portion of the electrical contacts between the different functional conductive elements of the 3D NAND device. These contacts may be located between the bit lines and the semiconductor channels (referred to as “contacts” in the present description). These contacts may also be located between the power supply lines and the word lines (referred to as “peripheral contacts” in the present description).
According to a third example, the 3D NAND memory obtained according to the method of the invention comprises a semiconductor substrate defining a horizontal plane, at least one semiconductor channel arranged along a vertical axis, and at least one bit line comprising a nickel or cobalt alloy deposited according to the metallization process of the invention.
The features of the metallization process described above, including the features of implementing the step of activating with a noble metal and the step of contacting the activated substrate with an electroless solution, apply to all three embodiments of the method according to the invention.
The fabrication method according to the invention may comprise, in addition to the metallization process, at least one other step. For example, the method of the invention may comprise, prior to the steps of activation and electroless deposition of the nickel or cobalt alloy, a step of depositing a layer of a dielectric material followed by a step of etching cavities in the dielectric material by photolithography. The walls of the cavities of a dielectric nature are then activated by the noble metal and metallized with the nickel or boron alloy as described above.
The invention relates to a semiconductor device obtainable by implementing the method described above.
Indeed, the metallization method of the invention can be used for the manufacture of three-dimensional semiconductor devices, such as three-dimensional integrated circuits or 3D NAND type storage memories, at the level of the creation of conductive lines, or such as V-DRAM storage memories or MIM capacitors forming part of the structure of DRAM storage memories.
The metallization method of the invention may for example be used for the manufacture of copper conductive lines comprising a step of depositing a nickel alloy layer on the walls of mineral oxide cavities, intended to be filled with copper in a subsequent step. These cavities have, for example, an average diameter at the opening ranging from 10 nm to 30 nm and a depth ranging from 20 nm to 100 nm. The average diameter at the opening of the cavities is preferably less than 500 nm, for example less than a value selected from the group consisting of 400 nm, 300 nm, 200 nm, 100 nm, 50 nm and 10 nm.
The invention is illustrated by the following examples. The temperature is between 20° C. and 25° C. unless otherwise stated. The pressure is equal to the atmospheric pressure.
In this example, the substrate used was a SiO2 coupon having dimensions of 4 cm×4 cm on each side and a thickness of 750 μm, and having vertical cavities with an aperture of about 10 μm and a height of about 100 micrometers.
The coupon was cleaned according to the chemical nature of the substrate. After this cleaning step, the coupon is rinsed thoroughly with deionized water, immersed in a beaker filled with deionized water subjected to ultrasound (40 kHz) for 2 minutes.
First stage of activation: In a beaker, 350 microliters of 3-[2-(2-aminoethylamino) ethylamino]propyl-trimethoxysilane and 15 mg of (ethylenediamine) palladium (II) chloride are dissolved in 80 ml of dry DMSO (maximum 50 ppm H2O), to obtain a first activation solution comprising 7.9·10−4 M of palladium complex and 24.4 mM of silane compound. The coupon prepared in step a1) is immersed in the beaker comprising the first activation solution at a temperature of 65° C. and under ultrasound for 15 minutes. The coupon is removed from the first activation solution, rinsed thoroughly with deionized water, and immersed in a beaker filled with deionized water subjected to ultrasound (40 kHz) for 30 seconds. The coupon is removed from the beaker, and rinsed thoroughly with deionized water.
Second stage of activation: The coupon was soaked in a solution of H2PdCl4 at a concentration of 10−3 mol/L at pH of 2.0 for 30 seconds. The coupon is removed from the solution, rinsed thoroughly with deionized water, immersed in a beaker filled with deionized water subjected to ultrasound (40 kHz) for 30 seconds. The coupon is then removed from the beaker, and rinsed thoroughly with deionized water.
In a container of 1 liter and a minimum of deionized water are introduced in order, 31.11 g of nickel sulphate hexahydrate (0.118 moles), 44.67 g of citric acid (0.232 moles), 52.26 g of N-methyl aminoethanol (0.700 moles) and 2.5 ppm of polyethyleneimine (PEI) (Mn=600 g/mol). The final pH was adjusted to 9.3 with N-methyl-aminoethanol and the total volume was adjusted to 1 liter with deionized water. To nine volumes of the previous solution, just before the next step, a volume of a reducing solution is added. The latter comprises 28 g/L of dimethyl-amino borane (DMAB; 0.475 moles) and 60.00 g of N-methyl aminoethanol (0.798 moles).
A nickel-boron alloy layer was deposited on the surface of the substrate treated in step b) by first immersing it in a beaker of deionized water and then soaking it in the electroless solution previously prepared and heated to 65° C., for a period ranging from 1 to 6 minutes. A gray and shiny metal cover is then observed on the coupon. The coupon is removed from the solution, rinsed thoroughly with deionized water, immersed in a beaker filled with deionized water subjected to ultrasound (40 kHz) for 30 seconds. The coupon is then removed from the beaker, rinsed thoroughly with deionized water and dried under a stream of nitrogen. The coupon is subject to Rapid Thermal Annealing (RTA) at 400° C. for ten minutes in a reducing atmosphere (4% hydrogen in nitrogen). The operation can be performed with a tubular oven or a hot plate.
The vertical cavities are covered with a thin continuous layer of nickel-boron alloy having a thickness measured under a microscope at a magnification of 150,000. Layer continuity, i.e. the thickness at which NiB grains become joined and at which conductivity is then measurable, is obtained with a deposition process lasting 2 minutes. In this case is measured a thickness of 7 nm and a resistivity of 110 μohm·cm. The resistivity was measured with the reference device (Automatic four point probe meter MODEL 280, available from Four Dynamics Inc.).
Example 1 was repeated except that the activation of the surface of the cavities included on the First Stage Activation (i.e., the Second Stage Activation was not completed).
After thermal annealing of the coupon obtained at the end of step c2), it is observed that all the vertical cavities are covered with a thin layer of the nickel-boron alloy of 18 to 40 nm thickness depending on the deposition time. The layer continuity, i.e. the thickness at which nickel-boron alloy metal grains become jointed and at which the conductivity is then measurable, is obtained after a deposition time of 3.5 min. In this case is measured a thickness of 18 nm and a resistivity is 330 μohm·cm.
The results of the resistivity measurements performed in Example 1 and Comparative Example 2 are presented in Table 1.
The results demonstrate that at very low thicknesses, the two stage activation process described herein produces a deposit exhibiting good conductivity, which is not achievable by a comparative process that only uses a single activation stage.
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/IB2023/000199 | 3/30/2023 | WO |
| Number | Date | Country | |
|---|---|---|---|
| 63325169 | Mar 2022 | US |
