Semiconductor device dielectric interface layer

Abstract

Embodiments related to methods for forming a film stack on a substrate are provided. One example method comprises exposing the substrate to an activated oxygen species and converting an exposed surface of the substrate into a continuous monolayer of a first dielectric material. The example method also includes forming a second dielectric material on the continuous monolayer of the first dielectric material without exposing the substrate to an air break.

Description

BACKGROUND

Some semiconductor devices employ a high-K dielectric material (relative to SiO₂) as a gate dielectric to achieve a desired equivalent oxide thickness (EOT) for the device. The high-K dielectric material may provide a film that is sufficiently thick to avoid undesirable leakage while providing other desirable electrical properties. However, forming a high-K dielectric material directly on some substrates may lead to interactions between the substrate and the high-K dielectric material that may undesirably alter device performance.

SUMMARY

Various embodiments are disclosed herein that relate to methods for forming a film stack on a substrate. One example method comprises exposing the substrate to an activated oxygen species and converting an exposed surface of the substrate into a continuous monolayer of a first dielectric material. The example method also includes forming a second dielectric material on the continuous monolayer of the first dielectric material without exposing the substrate to an air break.

This Summary is provided to introduce a selection of concepts in a simplified form that are further described below in the Detailed Description. This Summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to limit the scope of the claimed subject matter. Furthermore, the claimed subject matter is not limited to implementations that solve any or all disadvantages noted in any part of this disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 shows a flowchart illustrating a portion of a method for forming a film stack on a substrate without an air break according to an embodiment of the present disclosure.

FIG. 2 schematically shows a film stack including a continuous monolayer of a first dielectric material covered by a second dielectric material according to an embodiment of the present disclosure.

FIG. 3A shows a flowchart illustrating a portion of a method for determining a conversion time associated with converting an exposed surface of a substrate into a continuous monolayer of a first dielectric material according to an embodiment of the present disclosure.

FIG. 3B shows a flowchart illustrating another portion of the method shown in FIG. 3A.

FIG. 4 shows ellipsometer data for film stacks formed according to an embodiment of the present disclosure.

FIG. 5 shows ellipsometer data for film stacks formed according to an embodiment of the present disclosure.

FIG. 6 schematically shows an example semiconductor processing tool according to an embodiment of the present disclosure.

DETAILED DESCRIPTION

Modern semiconductor devices may include high-K gate dielectric materials to improve device performance relative to devices employing a SiO₂ gate dielectric at a given architectural size. The thickness of the gate dielectric material may shrink with the transistor size to provide a desired gate capacitance and/or drive current. The physical thickness of a SiO₂ gate dielectric may eventually become small enough so that leakage from the gate to the channel (e.g., from tunneling effects) raises power dissipation in the device to unacceptably high levels.

One approach to mitigate this effect is to replace the SiO₂ gate dielectric with a gate dielectric material having a dielectric constant value (K) that is relatively higher than the dielectric constant for SiO₂ (K˜3-4). For example, some semiconductor devices employ HfO₂ (K˜20-25), which may provide an equivalent oxide thickness (EOT) to a physically thinner layer of SiO₂ In turn, a gate including a high-K gate dielectric material may provide similar device turn-on characteristics while mitigating gate leakage and power dissipation.

As used herein, a high-K dielectric material refers to any suitable dielectric material having a dielectric constant greater than SiO₂. Accordingly, some transition metal oxides such as HfO₂ and TiO₂ (K˜50) may be considered high-K dielectric materials. Other example high-K dielectric materials include, but are not limited to HfO_x, TiO_x, SrO_x, ZrO_x, and LaO_x. Some non-transition metal-including oxides such as Al₂O₃ (K˜8) may be considered high-K dielectric materials. MgO is another non-limiting example of a non-transition metal-including oxide. Some high-K dielectric materials may have a dielectric constant of 200 or more (e.g., ultra-high-K dielectric materials), such as some alkaline earth metal-including metal oxides. For example, SrTi_xO_y, BaTi_xO_y, and Sr_xBa_(1-x)Ti_yO_z may be considered high-K dielectric materials.

However, forming a high-K dielectric material directly on some substrates may lead to undesirable reaction between the substrate and the high-K dielectric material or one or more precursors used to form the high-K dielectric material. For example, a layer of HfO₂ formed directly on a silicon substrate may form a hafnium silicide, potentially degrading the intended performance character of the device. One approach to managing interactions between high-K materials and underlying substrates is to form a protective oxide barrier layer on the substrate prior to deposition of the high-K material. For example, an interlayer oxide may be formed on the substrate between the substrate and the high-K gate dielectric to restrict or prevent diffusion of metal cations from the high-K material or from a precursor used to form the high-K gate dielectric material into the substrate.

However, some of the methods of forming interlayer oxides rely on native oxide growth processes that can be difficult to control. For example, one approach for forming a chemical oxide of SiO₂ on a silicon substrate uses an HF wet bath to strip the substrate back to bare silicon. The chemical oxide is then formed according to a “Standard Clean 1” or “SCI” process using a wet bath comprising NH₄OH, H₂O₂, and H₂O. Diffusion boundary layers formed within the liquid near the substrate surface may make it difficult to control the thickness of an oxide formed using a wet bath short of an equilibrium thickness that may be undesirably thick. For example, some chemical oxides may have a thickness approaching 7 to 8 Å, a thickness that may contribute at an undesirable level to the EOT of a device including a high-K gate dielectric.

Accordingly, the embodiments described herein are related to methods for forming a film stack on a substrate. One example method comprises exposing the substrate to an activated oxygen species and converting an exposed surface of the substrate into a continuous monolayer of a first dielectric material. The example method also includes forming a second dielectric material on the continuous monolayer of the first dielectric material without exposing the substrate to an air break.

The embodiments described herein are also related to methods for determining a conversion time associated with converting an exposed surface of a substrate into a continuous monolayer of a first dielectric material. One example method includes converting the exposed surface of the substrate into the first dielectric material via reaction of an activated oxygen species with the exposed surface of the substrate and then forming a second dielectric material on the first dielectric material without an air break. The example method also includes measuring a thickness of the second dielectric material and determining that a continuous monolayer of the first dielectric material is formed on the exposed surface of the substrate if the thickness matches a preselected thickness.

The embodiments described herein are also related to film stacks used for semiconductor device gate assemblies. In one example, a gate assembly for a semiconductor device comprises a continuous monolayer of a first dielectric material formed from an exposed surface of a semiconducting substrate via reaction of the exposed surface with an activated oxygen species and a high-K dielectric material formed upon the first dielectric material without an intervening air break.

Forming a high-K dielectric material on the first dielectric material without an air break may avoid spontaneous oxidation caused by exposure to ambient conditions that may undesirably thicken the EOT of the device. Such spontaneous oxidation processes may be uncontrolled, potentially causing device performance to vary across the substrate and/or the device. Further, avoiding an air break during processing may also prevent the introduction of contaminants and/or small particles within the device that may degrade performance.

Moreover, by controlling formation of the first dielectric material to a continuous monolayer, the contribution to the device's EOT from the first dielectric material may be managed with more precision than boundary layers formed from native growth processes, such as ambient air exposure and/or chemical oxide formation in a wet bath. For example, in some embodiments, formation of a continuous monolayer of SiO₂ on a silicon substrate as described herein may provide an SiO2 film that is physically less than or equal to 6 Å thick. Further, in some embodiments disclosed herein, a gate assembly formed from a film stack comprising a continuous monolayer of SiO₂ formed from a silicon using an oxygen plasma and an HfO₂ high-K dielectric film formed thereon without an air break may reduce device EOT by approximately 20% relative to devices having the same physical thickness of HfO₂ (within an acceptable tolerance) deposited on an SiO₂ layer formed via a chemical oxide generated by an HF strip followed by an SC 1 wet bath.

FIG. 1 shows a flowchart illustrating an embodiment of a method 100 for forming a film stack including a continuous monolayer of a first dielectric material and a second dielectric material formed on the first dielectric material without an air break. It will be appreciated that embodiments of method 100 may be performed by any suitable hardware and software, including the hardware and software described herein. It will be appreciated that portions of the processes described in method 100 may be omitted, reordered, and/or supplemented without departing from the scope of the present disclosure.

Method 100 comprises, at 102, supporting a substrate in a reactor. A first dielectric material may be formed on an exposed surface of any suitable substrate without departing from the present disclosure. In some embodiments, supporting the substrate in the reactor may include adjusting one or more reactor conditions, such as temperature, pressure, plasma processing conditions, and/or inert gas (e.g., Ar, N₂, or He) flow rate, to conditions suitable for film formation prior to processing the substrate, though it will be appreciated that such adjustments may be made at any suitable point during substrate processing. It will also be appreciated that such film formation conditions may vary according to film process chemistry, substrate surface termination, and so on. Further, in some embodiments, reactor conditions may be adjusted to avoid undesirable reactions in the process environment and/or on the substrate. For example, reactor conditions may be adjusted to avoid gas phase decomposition reactions for one or more of precursors and/or reactants, potentially avoiding film contamination from decomposition products and/or poor step coverage resulting from diffusion effects. Further, in some scenarios, reactor conditions may be adjusted to avoid condensation of precursors and/or reactants on various reactor surfaces, potentially avoiding small particle defect generation processes.

For example, Table 1, presented below, provides non-limiting examples for process conditions suitable for forming a first dielectric layer of SiO₂ on and/or within a surface of a silicon substrate and for forming a second dielectric layer of HfO₂ on the first dielectric layer without an intervening air break.

	TABLE 1
	Lower	Middle	Upper
	Range	Range	Range
	Example	Example	Example
First dielectric material conversion
process: Si converted
to SiO2 using an oxygen plasma
Temperature (° C.)	25	50	400
Total pressure (Torr)	0.5	1	5
Plasma power (W)	10	50	200
Oxygen flow rate (sccm)	10	100	1000
Diluent flow rate (sccm)	100	500	3000
Conversion Time (sec)	0.1	1	60
Second dielectric material conversion
process: HfO2 deposited using
atomic layer deposition
Temperature (° C.)	200	300	400
Total pressure (Torr)	1	4	9
Dosing times for precursor (HfCl4)	0.1	2	10
and reactant (H2O)

At 104, method 100 comprises supplying an exposed substrate surface with activated oxygen species. As used herein, an activated oxygen species generally refers to a chemically active state of oxygen generated from an oxygen-including reactant. In some embodiments, activated oxygen species may be provided by an ozone source. For example, ozone may be provided by an ozone generator supplied with oxygen. In some embodiments, activated oxygen species may be provided by an oxygen plasma source. In one scenario, activated oxygen ions may generated from oxygen gas by plasma formed proximate to the substrate (e.g., direct plasma). In another scenario, the activated oxygen ions may be supplied from an oxygen plasma generated away from the substrate surface (e.g., remote plasma). It will be appreciated that selection of a source for generating an activated oxygen species may depend on one or more application-related criteria, such as sensitivity of the substrate to ion bombardment, a desired flux of activated oxygen species at the substrate surface, a desired penetration depth of the activated oxygen species within the substrate surface, and so on. Such considerations or others may also affect selection of the activated oxygen species.

While the description above generally relates to the conversion of an exposed substrate surface into a layer of an oxygen-including dielectric material, it will be appreciated that any other suitable material may also be included in the first dielectric material without departing from the scope of the present disclosure.

For example, in some embodiments, nitrogen may also be included in the first dielectric material to adjust one or more electrical and/or physical characteristics of the first dielectric material, such as the dielectric constant, etch rate, and so on. Accordingly, in some embodiments, supplying the exposed surface with activated oxygen species at 104 may include supplying the exposed surface with an activated nitrogen species. As used herein, an activated nitrogen species generally refers a chemically active state of nitrogen generated from a nitrogen-including reactant. In some embodiments, activated nitrogen species may be provided by a nitrogen plasma source. In one scenario, activated nitrogen ions may generated from nitrogen gas by direct plasma, while, in another scenario, activated nitrogen ions may be supplied by remote plasma.

It will be appreciated that, if supplied, an activated nitrogen species may be provided by a source that is different than a source used to supply activated oxygen species in some embodiments, though in some instances the activated nitrogen and oxygen species may be generated by a common generation source. Further, it will be appreciated that selection of a source for generating an activated nitrogen species may depend on one or more of the application-related criteria noted above or on other suitable factors. Such considerations or others may also affect selection of the activated nitrogen species.

At 106, method 100 comprises converting an exposed surface of the substrate into a continuous monolayer of the first dielectric material. The activated oxygen species, and in some embodiments, the activated nitrogen species, reacts with the surface of the substrate material to form the first dielectric material. The conversion reaction transforms the exposed surface of the substrate into the first dielectric material. Put differently, the activated species consumes a portion of the exposed surface of the substrate to generate the first dielectric material. For example, in some embodiments, activated oxygen species may react with exposed surfaces of a silicon substrate to convert those surfaces into suitable silicon oxides (SiO_x), such as SiO₂. In some other embodiments, activated oxygen and nitrogen species may react with exposed surfaces of a silicon substrate to convert those surfaces into suitable silicon oxynitrides (SiO_xN_y), such as SiON.

The conversion reaction is performed until a continuous monolayer of the first dielectric material is formed. As used herein, a continuous monolayer refers to an uninterrupted single layer, within an acceptable tolerance, of a specified material, such as the first dielectric material. Without wishing to be bound by theory, activated species supplied to the substrate may react with the substrate to form islands or domains of the first dielectric material. Continued supply of the activated species may cause new domains to form and existing domains to grow until the domains merge together to form a continuous monolayer.

It will be appreciated that the physical characteristics of a monolayer may depend upon the material being formed. For example, a SiO₂ film formed within a silicon substrate may have a monolayer thickness of approximately 4 Å based upon the bond length and unit cell dimensions for SiO₂. Other dielectric materials formed from silicon substrates, such as silicon suboxides and silicon oxynitrides may have different monolayer thicknesses. Dielectric materials formed from other substrate materials may also have different monolayer thicknesses that are specific to the selected material systems.

Because the conversion process described above consumes a portion of the substrate, it will be appreciated that activated species may diffuse through the first dielectric material to reach unreacted substrate. If several layers of dielectric material accumulate, these diffusional effects may slow or halt the conversion of the substrate material. However, because several layers of dielectric material may compromise device performance, the supply of active species is managed so that the supply is curtailed after generating a continuous monolayer of the dielectric material. It will be appreciated that the supply may be controlled in any suitable manner without departing from the scope of the present disclosure. In some embodiments, activated species may be supplied only during a predetermined conversion time. Selection of the predetermined conversion time is discussed in more detail below with respect to FIGS. 3A and 3B.

At 108, method 100 comprises forming a second dielectric material on the first dielectric material without an air break. The second dielectric material may be formed on the first dielectric material in any suitable manner. For example, the second dielectric material may be formed on the first dielectric material by a thermal and/or plasma moderated deposition technique (e.g., atomic layer deposition or chemical vapor deposition). As one example of an atomic layer deposition technique, the second dielectric material may be deposited on the first dielectric material by adsorbing a dielectric material precursor on the continuous monolayer of the first dielectric material. A layer of the second dielectric material is then formed from interaction of the adsorbed precursor with a subsequently introduced reactant. Repeating the sequence builds additional layers of second dielectric material.

Forming the second dielectric material on the first dielectric material without an air break may avoid oxide growth caused by exposure of the first dielectric material to oxygen and/or humidity present at ambient conditions. In some settings where the first dielectric material is used as a liner or barrier layer prior to formation of a high-K dielectric, spontaneous oxidation caused by exposure to ambient conditions may undesirably thicken the EOT of the device. Moreover, because such spontaneous oxidation processes are relatively uncontrolled, the first dielectric material may vary in thickness across the substrate and/or the device. Further, forming the second dielectric material on the first dielectric material without an air break may also avoid exposing the surface of the first dielectric material to ambient contaminants and/or small particles that may interfere with subsequent processing.

Accordingly, in some embodiments, the first dielectric material may be formed in a first processing module of a semiconductor processing tool. The substrate may then be moved, without exposure to air or ambient humidity, to a second processing module included in the semiconductor processing tool for formation of the second dielectric material.

In some embodiments, forming the second dielectric material on the first dielectric material without an air break may include forming the second dielectric material on the first dielectric material without a vacuum break. In some of such embodiments, the first dielectric material may be formed in a first processing module of a semiconductor processing tool. The substrate may then be moved, without a vacuum break, to a second processing module included in the semiconductor processing tool for formation of the second dielectric material. In some other of such embodiments, the second dielectric material may be formed in the same processing module as the first dielectric material.

FIG. 2 schematically shows an embodiment of a film stack 200 that may be used to form a gate assembly for a semiconductor device. Film stack 200 may be formed by the methods disclosed herein or by any suitable method. Film stack 200 comprises a semiconducting substrate 202, a first dielectric material 204, and a second dielectric material 206.

It will be appreciated that semiconducting substrate 202 may include any suitable semiconducting substrate without departing from the scope of the present disclosure. Non-limiting examples of semiconducting substrate 202 include Si, Ge, SiGe, GaAs, InGaAs, and InP. In some embodiments, semiconducting substrate 202 may be formed as a suitable wafer, such as a 300-mm double-sided polished wafer of a suitable material. Further, semiconducting substrate 202 may have any suitable topography, including planar and non-planar surfaces.

First dielectric material 204 is a continuous monolayer formed by reaction of an exposed surface of semiconducting substrate 202 with an activated oxygen species. For example, in some embodiments, activated oxygen species may react with exposed surfaces of a silicon substrate to convert those surfaces into suitable silicon oxides (SiO_x), such as SiO₂. In some other embodiments, activated oxygen and nitrogen species may react with exposed surfaces of a silicon substrate to convert those surfaces into suitable silicon oxynitrides (SiO_xN_y), such as SiON. It will be appreciated that other suitable first dielectric materials may be formed from other substrate materials. For example, in embodiments where activated oxygen species is supplied to a Ge, SiGe, GaAs, InGaAs, or an InP substrate, first dielectric material 204 comprises a continuous monolayer of GeO_x, SiGeO_x, GaAsO_x, InGaAsO_x, or InPO_x, respectively. In embodiments where activated oxygen and nitrogen species are supplied to a Ge, SiGe, GaAs, InGaAs, or an InP substrate, first dielectric material 204 comprises a continuous monolayer of GeO_xN_y, SiGeO_xN_y, GaAsO_xN_y, InGaAsO_xN_y, or InPO_xN_y, respectively.

Second dielectric material 206 is a dielectric material formed on first dielectric material 204 without an intervening air break. In some embodiments, second dielectric material 206 may include a binary or a tertiary transition metal oxide. In some embodiments, second dielectric material 206 may include a high-K dielectric material. Non-limiting examples of second dielectric material 206 include HfO_x, TiO_x, SrO_x, ZrO_x, LaO_x, SrTi_xO_y, AlO_x, MgO, BaTi_xO_y, and Sr_xBa_(1-x)Ti_yO_z.

While not shown in FIG. 2, a suitable gate electrode may be formed on second dielectric material 206, so that first dielectric material 204 and second dielectric material 206 act as a gate dielectric for a gate assembly that includes the gate electrode. It will be appreciated that film stack 200 may be formed into a gate assembly for a semiconductor device by suitable patterning steps, potentially including one or more lithographic, etch, and mask steps. Other structures, including spacers, contacts, source and drain regions, and so on may also be included in a semiconductor device including a gate assembly that includes film stack 200.

As introduced above, a continuous monolayer of the first dielectric is formed by converting at least enough of the substrate to form the continuous monolayer while curtailing conversion thereafter. FIGS. 3A and 3B show a flowchart that illustrates an embodiment of a method 300 for determining a conversion time associated with converting an exposed surface of a silicon substrate into a continuous monolayer of a first dielectric material comprising silicon dioxide according to an embodiment of the present disclosure. It will be appreciated that embodiments of method 300 may be performed by any suitable hardware and software, including the hardware and software described herein. It will be appreciated that portions of the processes described in method 300 may be omitted, reordered, and/or supplemented without departing from the scope of the present disclosure. For example, it will be appreciated that various decision structures included in method 300 may be altered, omitted, or supplemented in some embodiments involving other substrate materials, other first dielectric materials, and/or other second dielectric materials.

At 302, method 300 comprises supplying an active species to exposed surfaces of a silicon substrate for a selected time to form SiO₂. The time selected for converting the substrate to the SiO₂ may be any suitable time, which may be adjusted as described in more detail below. While it will be appreciated that any suitable process variable such as temperature, pressure, partial pressure, plasma power, and so on may be adjusted during conversion 302, it is preferable that those process parameters be selected in advance. In turn, method 300 may be focused on time selection for a particular substrate material conversion process, which may expedite time selection and overall process development.

At 304, method 300 comprises forming HfO₂ on the SiO₂ without an air break. Forming the second dielectric material at 304 includes processing the substrate under conditions suited to form a predetermined amount or thickness of the HfO₂ layer on the SiO₂ layer. In some of such embodiments, the substrate may be processed for a predetermined film formation time. For example, in some embodiments where the second dielectric material is deposited by a thermal- or plasma-based chemical vapor deposition technique, the substrate may be processed for a predetermined deposition time. In some embodiments where the second dielectric material is deposited by an atomic layer deposition technique, the substrate may be processed for a predetermined number of film formation cycles.

Processing the substrate so as to form a predetermined amount of the second dielectric material may provide an approach to identifying whether the SiO₂ film was sufficiently formed into a closed, continuous monolayer of film. In some settings, a precursor used to form the second dielectric material may adsorb at a different rate on the substrate than on the SiO₂. Relative differences in adsorption rate and/or extent of adsorption may be due to differences in the chemical termination of those surfaces.

For example, HfO₂ may be deposited using atomic layer deposition by sequentially supplying a substrate with separate, alternating doses or pulses of a hafnium-including precursor and an oxygen-including reactant. Reactions 1 and 2 illustrate a simplified form of the two processes that make up a sequential cycle used to form a layer of HfO₂ using atomic layer deposition.HfCl_{4 (gas)}+surface active site→HfCl_{3 (adsorbed)}+Cl_(adsorbed)  Reaction 1:2H₂O_(gas)+HfCl_{3 (adsorbed)}+Cl_(adsorbed)→HfO₂+4HCl_(gas)  Reaction 2:

It will be appreciated that not all sites on the substrate surface may be active for adsorbing precursor. In some settings, the chemical termination of the surface may affect the population of surface active sites. For example, a surface terminated with hydrogen may be less reactive to some processes than a surface terminated with hydroxyl groups. Without wishing to be bound by theory, hydrogen-terminated surfaces may have fewer dangling bonds relative to other surface substituent groups, potentially providing a lower energy and less reactive surface. Accordingly, some surfaces may undergo a surface transformation before the surface is active for film formation. For example, water may be split to transform the surface from a hydrogen-terminated state into a hydroxyl-terminated state. It may take one or more sequential cycles to nucleate surface active sites suited for precursor adsorption and film deposition. Meanwhile, little film formation may occur during surface active site nucleation.

The delay associated with preparing the surface for subsequent film formation may be referred to as a nucleation phase. It will be appreciated that the duration of the nucleation phase, if any exists, may vary according to the identities of the substrate and the precursor. The nucleation phase duration may also vary according to an identity of the first dielectric material and the extent of conversion of the substrate surface to the first dielectric material. As introduced above, in some settings, a precursor used to form the second dielectric material may adsorb at a different rate on the substrate than on the first dielectric material. For example, a silicon substrate may be hydrogen terminated while a silicon oxide surface formed from the silicon substrate may be terminated with a non-hydrogen species, such as a surface hydroxyl. In some of such examples, a precursor used for forming the second dielectric material, such as HfCl₄, may have a higher sticking probability on the silicon oxide surface than on the silicon surface. The difference in sticking probability may lead to a difference in a likelihood of forming the second dielectric material, as a precursor that is more likely to adsorb is may be more likely to participate in subsequent film-forming reactions.

For example, FIG. 4 shows ellipsometer data 400 for film stacks formed according to an embodiment of the methods described herein comprising a layer of HfO₂ formed on top of a layer of SiO₂ grown from a silicon substrate using an oxygen plasma. The SiO₂ layers represented in FIG. 4 have different thicknesses generated by varying the exposure of the silicon substrate to oxygen radicals and/or ions generated by direct oxygen plasma prior to deposition of HfO₂. An identical number of atomic layer deposition cycles were used when forming the HfO₂ layers. However, as explained in more detail below, the resulting thickness of HfO₂ may vary according to an extent of conversion of the silicon surface to SiO₂.

Ellipsometer data 400 includes film thickness data 402 and film non-uniformity data 404 collected as a function of oxygen plasma exposure time. Ellipsometer data 400 represents data averaged from a suitable number of measurement sites on the substrates. Thickness data 402 was collected from 49 measurement sites arranged on each substrate surface, though any suitable number of sites may be employed without departing from the scope of the present disclosure. Non-uniformity data 404 represents within-substrate non-uniformity, which was calculated by dividing the standard deviation of the measured thickness data by the average thickness. The calculated non-uniformity values are expressed as percentages by multiplying the results by 100.

For illustrative purposes, ellipsometer data 400 is divided into Region A and Region B. Region A illustrates a portion of ellipsometer data 400 roughly corresponding to a silicon surface that exhibits no conversion to SiO₂ or conversion to less than a continuous monolayer of SiO₂. Consequently, some of the atomic layer deposition cycles are used transforming the surface in a nucleation phase. Only the remaining atomic layer depositions cycles are able to deposit HfO₂ film on the surface. Because the total film deposited during the HfO₂ formation process in Region A is less than an amount that would ordinarily be deposited by an identical number of atomic layer deposition cycles, the total thickness shown in thickness data 402 in Region A is less than expected. Moreover, because the HfO₂ process begins with a nucleation process, islands of HfO₂ formed on the surface from nucleation sites may also lead to a surface that has a greater non-uniformity than expected, as shown in non-uniformity data 404.

Region B illustrates a portion of ellipsometer data 400 roughly corresponding to a silicon substrate exhibiting conversion to a continuous monolayer of SiO₂ prior to formation of HfO₂. Because the SiO₂ surface may be more susceptible to HfO₂ formation than a silicon surface, fewer atomic layer deposition cycles may be used in a nucleation phase in Region B than in Region A. In turn, more deposition cycles may be available for forming layers of HfO₂, so that the total thickness approaches an amount of HfO₂ that is expected to be deposited by those atomic layer deposition cycles. In settings where there is virtually no nucleation phase for forming HfO₂ on the SiO₂ surface, the thickness of HfO₂ formed may approach a theoretical value for atomic layer deposition. Further, because the HfO₂ films are formed on a continuous monolayer of film, the total film stack exhibits a more uniform surface relative to surfaces having discontinuous, sub-monolayer films of SiO₂ beneath the HfO₂, as shown in non-uniformity data 404.

In the example shown in FIG. 4, the duration of oxygen plasma corresponding to transition 406 represents the conversion time to form a continuous monolayer of silicon oxide from the silicon substrate for these conversion conditions. While other conversion conditions and/or material systems may have different conversion times, it will be appreciated that the approach described herein may also be employed to identify a conversion time for such conditions and/or systems. It will also be appreciated that extending the plasma duration may increase the thickness of the silicon oxide layer in some settings. In settings where the silicon oxide layer is included in a gate assembly, increasing the thickness of the silicon oxide layer may increase the EOT of the gate.

FIG. 5 shows ellipsometer data 500 for film stacks formed according to an embodiment of the methods described herein comprising a layer of HfO₂ formed on top of a layer of SiO₂ grown from a silicon substrate using ozone. Thickness data 502 and non-uniformity data 504 exhibit behavior like that described above regarding FIG. 4. In the example shown in FIG. 5, transition 506 indicates that a continuous monolayer of SiO₂ is formed from the silicon substrate after approximately 5 seconds of ozone exposure under those conversion conditions.

Returning to FIG. 3A, at 306, method 300 comprises measuring a thickness of a film stack comprising the SiO₂ and HfO₂ films. The measured value provides a basis for determining whether the SiO₂ exists in a continuous monolayer, as explained in more detail below. It will be appreciated that any suitable technique for measuring the HfO₂ may be employed without departing from the scope of the present disclosure. For example, thickness may be measured by an optical technique (e.g., ellipsometry), by a microscopy technique (e.g., transmission electron microscopy), by an electron spectroscopy technique (e.g., X-ray photoelectron spectroscopy), and so on.

At 308, method 300 comprises determining whether the thickness of the film stack is less than expected based upon the HfO₂ formation conditions. In turn, it may be possible to determine whether a continuous monolayer of SiO₂ is formed. If the thickness of the film stack is less than expected, method 300 continues to 310 in FIG. 3B. At 310, it is determined that, because the thickness of the film stack (or in some embodiments, the thickness of the HfO₂ layer) is less expected for the given HfO₂ formation conditions, an incomplete or discontinuous monolayer of the first dielectric material may have been formed.

Without wishing to be bound by theory, a thinner than expected HfO₂ layer may result when an HfO₂ precursor or reactant has a higher probability of sticking on the SiO₂ than on the silicon substrate. Initial formation of the HfO₂ layer on the silicon substrate may be inhibited by a nucleation phase that is longer than a nucleation phase associated with forming the HfO₂ layer on SiO₂, causing the HfO₂ layer to be thinner than expected. Consequently, at 312, method 300 includes increasing the selected conversion time and repeating method 300. Increasing the selected conversion time may provide additional time for the SiO₂ material to close into a continuous monolayer.

If the thickness of the film stack is not less than expected at 308, method 300 continues to 314 in FIG. 3B. At 314, method 300 comprises determining whether the thickness of the film stack is greater than expected, which may indicate that the SiO₂ layer thickness exceeds one monolayer. If the thickness of the film stack is not greater than expected, method 300 continues to 316, where it is determined that a continuous monolayer of SiO₂ was formed. In turn, the conversion time selected for conversion 302 may be used to form a continuous monolayer of SiO₂.

If the thickness of the film stack is greater than expected, method 300 continues to 318, where it is determined that more than one monolayer of SiO₂ was formed. Consequently, at 320, method 300 includes decreasing the selected conversion time and repeating method 300. Decreasing the selected conversion time may reduce the time available for the SiO₂ material to grow beyond a continuous monolayer.

It will be appreciated that the silicon/SiO₂ example disclosed above is merely illustrative, and that some portions of method 300 may vary according to materials- and/or process-related considerations. In some embodiments, determinations based on thickness comparisons may be related to the extent of conversion of the substrate into the first dielectric material may depend on a sticking probability of a precursor, reactant, and/or reaction product (e.g., a gas phase deposition product participating in a CVD reaction) associated with formation of the second dielectric material on the first dielectric material relative to the substrate. In some embodiments, determinations based on thickness comparisons may be related to variations in substrate and first dielectric material physical surface conditions (e.g., faceting and so on) and chemical surface conditions (e.g., the population of dangling bonds, surface adsorbed groups including hydrogen, hydroxyl, and so on).

For example, if a precursor used to form the second dielectric material has a higher sticking probability on the substrate than on the first dielectric material, the thickness of the second dielectric material may be greater than expected. In such embodiments, the greater-than-expected thickness may be used to increase the selected conversion time. Further, in some of such embodiments, the measured thickness of the second dielectric material, on its own, may be helpful in distinguishing between a film stack comprising an incomplete, discontinuous layer of first dielectric material and a film stack comprising more than one monolayer of the first dielectric material.

In some embodiments, the film stacks and structures described herein may be formed using a suitable semiconductor processing tool. FIG. 6 schematically shows a top view of an embodiment of a semiconductor processing tool 600 including a plurality of semiconductor processing modules 602. While the depicted embodiment includes two modules, it will be appreciated that any suitable number of semiconductor processing modules may be provided. For example, some processing tools may include just one module while other processing tools may include more than two modules.

FIG. 6 also shows load locks 604 for moving substrates between portions of semiconductor processing tool 600 that exhibit ambient atmospheric pressure conditions and portions of the tool that are at pressures lower than atmospheric conditions. An atmospheric transfer module 608, including an atmospheric substrate handling robot 610, moves substrates between load ports 606 and load locks 604, where a portion of the ambient pressure is removed by a vacuum source (not shown) or is restored by backfilling with a suitable gas, depending on whether substrates are being transferred into or out of the tool. Low-pressure substrate handling robot 612 moves substrates between load locks 604 and semiconductor processing modules 602 within low-pressure transfer module 614. Substrates may also be moved among the semiconductor processing modules 602 within low-pressure transfer module 614 using low-pressure substrate handling robot 612, so that sequential and/or parallel processing of substrates may be performed without exposure to air and/or without a vacuum break.

FIG. 6 also shows a user interface 620 connected to a system process controller 622. User interface 620 is adapted to receive user input to system process controller 622. User interface 620 may optionally include a display subsystem, and suitable user input devices such as keyboards, mice, control pads, and/or touch screens, for example, that are not shown in FIG. 6.

FIG. 6 shows an embodiment of a system process controller 622 provided for controlling semiconductor processing tool 600. System process controller 622 may operate process module control subsystems, such as gas control subsystems, pressure control subsystems, temperature control subsystems, electrical control subsystems, and mechanical control subsystems. Such control subsystems may receive various signals provided by sensors, relays, and controllers and make suitable adjustments in response.

System process controller 622 comprises a computing system that includes a data-holding subsystem 624 and a logic subsystem 626. Data-holding subsystem 624 may include one or more physical, non-transitory devices configured to hold data and/or instructions executable by logic subsystem 626 to implement the methods and processes described herein. Logic subsystem 626 may include one or more physical devices configured to execute one or more instructions stored in data-holding subsystem 624. Logic subsystem 826 may include one or more processors that are configured to execute software instructions.

In some embodiments, such instructions may control the execution of process recipes. Generally, a process recipe includes a sequential description of process parameters used to process a substrate, such parameters including, but not limited to, time, temperature, pressure, and concentration, as well as various parameters describing electrical, mechanical, and environmental aspects of the tool during substrate processing. The instructions may also control the execution of various maintenance recipes used during maintenance procedures.

In some embodiments, such instructions may be stored on removable computer-readable storage media 628, which may be used to store and/or transfer data and/or instructions executable to implement the methods and processes described herein, excluding a signal per se. It will be appreciated that any suitable removable computer-readable storage media 628 may be employed without departing from the scope of the present disclosure. Non-limiting examples include DVDs, CD-ROMs, floppy discs, and flash drives.

It is to be understood that the configurations and/or approaches described herein are exemplary in nature, and that these specific embodiments or examples are not to be considered in a limiting sense, because numerous variations are possible. The specific routines or methods described herein may represent one or more of any number of processing strategies. Thus, the various acts illustrated may be performed in the sequence illustrated, in other sequences, or omitted in some cases.

The subject matter of the present disclosure includes all novel and nonobvious combinations and subcombinations of the various processes, systems and configurations, and other features, functions, acts, and/or properties disclosed herein, as well as any and all equivalents thereof.

Claims

1. A method of forming a film stack on a substrate, the method comprising: determining a conversion time for converting an exposed surface of a substrate into a continuous monolayer;under vacuum conditions, exposing the substrate to a plasma activated oxygen species for the conversion time;converting an exposed surface of the substrate into a continuous monolayer of a first dielectric material; andforming a second dielectric material on the continuous monolayer of the first dielectric material without exposing the substrate to an air break,wherein the step of determining a conversion time comprises measuring a thickness of a film stack comprising the first dielectric material and the second dielectric material.

2. The method of claim 1, where converting the exposed surface into the continuous monolayer includes supplying the plasma activated oxygen species to the exposed surface until domains of the first dielectric material formed by reaction of the plasma activated oxygen species with the substrate merge into the continuous monolayer.

3. The method of claim 1, where forming the second dielectric material comprises depositing the second dielectric material by adsorbing a dielectric material precursor on the continuous monolayer of the first dielectric material.

4. The method of claim 1, where forming the second dielectric material without exposing the substrate to an air break comprises forming the second dielectric material without exposing the substrate to a vacuum break.

5. The method of claim 1, where exposing the substrate to the plasma activated oxygen species includes exposing the substrate to an activated nitrogen species.

6. The method of claim 1, where the substrate is selected from the set consisting of Si, Ge, SiGe, GaAs, InGaAs, and InP, and where the first dielectric material is selected from the set consisting of SiOxNy, SiOx, GeOxNy, GeOx, SiGeOxNy, SiGeOx, GaAsOxNy, GaAsOx, InGaAsOxNy, InGaAsOx, InPOxNy, and InPOx.

7. The method of claim 1, where the second dielectric material includes a high-K dielectric material.

8. The method of claim 7, wherein the high-K dielectric material has a dielectric constant greater than the dielectric constant of SiO2.

9. The method of claim 1, where the plasma activated oxygen species is provided by one or more of an ozone source and a remote oxygen plasma source.

10. The method of claim 1, where the substrate is hydrogen terminated and where the first dielectric material is non-hydrogen terminated.

11. The method of claim 1, where the step of converting an exposed surface of the substrate into a continuous monolayer comprises exposing the substrate to a vacuum pressure between 0.5 Torr and 5 Torr.

12. The method of claim 1, where the step of forming a second dielectric material comprises exposing the substrate to a vacuum pressure between 1 Torr and 9 Torr.

13. The method of claim 1, where the second dielectric is selected from the group consisting of HfOx, TiOx, SrOx, ZrOx, LaOx, SrTixOy, AlOx, MgO, BaTixOy, and SrxBa(1-x)TiyOz.

14. The method of claim 1, where the step of converting an exposed surface of the substrate and the step of forming a second dielectric material are performed in the same processing chamber.

15. The method of claim 1, where the step of converting an exposed surface of the substrate and the step of forming a second dielectric material are performed in different processing chambers.

16. The method of claim 1, wherein the film stack forms part of a semiconductor device gate assembly.

17. The method of claim 1, wherein a thickness of the continuous monolayer is less than or equal to 6 Angstroms.

18. The method of claim 1, wherein the plasma activated oxygen species is generated from oxygen gas.

19. The method of claim 1, wherein the plasma activated oxygen species is generated proximate the exposed surface of the substrate.

20. The method of claim 1, wherein the continuous monolayer comprises SiOxNy.