KR20130129149A - Conformal doping via plasma activated atomic layer deposition and conformal film deposition - Google Patents

Abstract

Methods for doping a substrate patterned in a reaction chamber are disclosed herein. The methods includes a step of forming a first conformal film layer having a dopant source including a dopant and a step of forming a conformal doping profile by driving a part of the dopant into the substrate. In several execution cases, the step of forming the first conformal film layer includes a step of inserting a dopant precursor into the reaction chamber, a step of absorbing the precursor under conditions for forming an absorption limit layer, and a step of making the dopant precursor react in order to form a dopant source. In addition, the reaction chamber, a gas inlet, and a controller having a machine-readable code are included as substrate doping devices and the controller includes commands for operating the gas inlet to absorb the dopant precursor by inserting the precursor and commands for making the absorbed precursor react to form a film layer containing the dopant source. The substrate doping devices are shown on the document. [Reference numerals] (AA) Dopant ALD cycle(s);(BB) Dopant precursor pulse and absorption;(CC) Pump to base and selective purge;(DD) (oxidizer pulse when plazma oxide is added);(EE) Add one of plazma :plazma oxide (for instance Ar + oxidizer + RF) or deactivation plazma (for instance He +RF);(FF) Pump to base and selective purge;(GG) Dopant ALD cycle;(HH) Selective additional ALd cycle (repeat G-J one or more time)

Description

CONFORMAL DOPING VIA PLASMA ACTIVATED ATOMIC LAYER DEPOSITION AND CONFORMAL FILM DEPOSITION}

At present, there is considerable interest in three-dimensional (3D) transistor structures such as multi-gate FinFETs and 3D memory devices. Examples of some 3D structures are described in “Tri-gate,” by J. Kavalieros (Intel), Symp. VLSI Tech, p. 50 (2006) and “Opportunities and Challenges of FinFET as a Device Structure Candidate for 14nm Node CMOS Technology,” ECS Transactions, Vol. 34, Issue 1, pp. 81-86 (2011), both of which are incorporated herein by reference for all and all purposes. Advanced 3D gate structures are employed at 22 nm technology nodes, and are likely to be employed at other technology nodes with feature sizes below 22 nm. Transistors with 3D gate structures typically employ source and drain regions formed in thin vertical structures. These vertical structures are difficult to dope using conventional ion implantation techniques, plasma doping techniques, and generally techniques involving the transport of ions under an electric field. The difficulty is generally evident when doping sidewalls, especially in high aspect ratio structures.

Thus, ion implantation processes are usually directional, and they may have many gate architecture designs, such as 3D gate designs and / or closely spaced fins that may require doping on lateral and vertical surfaces. There is a tendency to be unsuitable for the manufacture of gate architectures having. If an attempt is made to use ion implantation to dope these surfaces, such as by changing the implant tilt angle, this is especially true in the dopant dose retention and diffusion range, especially on the sidewalls. May result in significant variability. Frequently, in dense arrays of i3D structures, there may be shadowing effects on the directional ion beam in the injector, which causes serious dose retention problems for oblique implant angles. In addition, plasma doping techniques are of recent interest. However, there are two major drawbacks to this technique. (a) it may result in simultaneous sputter erosion, which may be due to high energy ions in the plasma, and (b) dopant dose and conformality are ions in the plasma. It appears to be sensitive to the ratio of radicals to radicals, which leads to difficulties in process control. Thus, there is a continuing need for better doping methods that may be used to fabricate many current gate architecture designs.

Disclosed herein are methods of doping a patterned substrate in a reaction chamber. In some embodiments, the present methods include a first conformal film comprising a dopant source comprising a dopant, directly on the surface of a patterned substrate having three-dimensional features or on a material adhered to the surface. Forming a layer, and driving a portion of the dopant from the first film layer into the substrate to form a conformal doping profile in the substrate. In some embodiments, forming the first film layer comprises introducing a dopant precursor into the reaction chamber, on the surface of the patterned substrate directly or under the conditions under which the dopant precursor forms an adsorption limiting layer. Adsorbing the dopant precursor on a material adhered to the surface, and reacting the adsorbed dopant precursor to form a dopant source.

 In some embodiments, forming the first film layer may include atomic layer deposition, and may also include removing the nonadsorbed dopant precursor from the reaction chamber. In certain such embodiments, the removal step may occur after the adsorption step but before the reaction step. In some embodiments, the removing step may include pumping the nonadsorbed dopant precursor from the reaction chamber by pumping the reaction chamber at a base pressure using a vacuum pump, while In some embodiments, the removing step may include purging the reaction chamber with an inert gas.

Some methods disclosed herein may further include forming a second conformal film layer containing a dielectric material, and some methods disclosed herein comprise a dopant comprising a dopant on a material adhered to the surface of the substrate. The method may further comprise forming a third conformal film layer comprising a source, wherein a substantial portion of the second film layer is intersperated between the first film layer and the third film layer. In certain such embodiments, the methods may further include driving a portion of the dopant from the third film layer into the substrate to form a conformal doping profile in the substrate. In some embodiments, forming the third film layer includes introducing a dopant precursor into the reaction chamber, over the material adhered to the surface under conditions such that the dopant precursor forms an adsorption-limited layer. Adsorbing the dopant precursor to and reacting the adsorbed dopant precursor to form a dopant source. In some embodiments, adsorbing the dopant precursor when forming the first film layer may include atomic layer deposition, and adsorbing the dopant precursor when forming the third film layer includes atomic layer deposition. You may. In some embodiments, forming the first film layer may include removing the nonadsorbed dopant precursor from the reaction chamber, and forming the third film layer comprises the nonadsorbed dopant from the reaction chamber. Removing the precursor. In some embodiments, during each formation of the first and third membrane layers, the removal step occurs after the adsorption step but before the reaction step. In some embodiments, the removal may include pumping the nonadsorbed dopant precursor from the reaction chamber by pumping the reaction chamber at a base pressure using a vacuum pump. In some embodiments, the removal may include purging the reaction chamber with an inert gas.

In some embodiments, reacting the adsorbed dopant precursor to form the first film layer may include activation using plasma, and reacting the adsorbed dopant precursor to form the third film layer Activation with plasma may also be included. In some embodiments, the plasma is an oxidative plasma. In some embodiments, the plasma is an inert plasma. In some embodiments, driving a portion of the dopant from the first film layer and the third film layer may include a thermal mediated diffusion process. In some embodiments, the thermal mediated diffusion process involves annealing. In some embodiments, the annealing is laser spike annealing.

In some embodiments, the dopant may comprise a valence III or V element. In some embodiments, the element is boron, phosphorus, germanium or arsenic. In some embodiments, the dopant is boron, the dopant precursor is alkyl borate and the dopant source is boron oxide or boron oxyhydride. In some embodiments, the alkyl borate is trimethyl borate. In some embodiments, the boron oxide is B ₂ O ₃ . In some embodiments, the dopant is arsenic, the dopant precursor is selected from alkylarcin, alkoxyarcin, and the aminoarcin chemical family, and the dopant source is arseno silicate, arsenic doped silicate glass, arsenic oxide Or arsenic oxyhydride. In some embodiments, the dopant precursor is arsine, triethylarsenate, trimethylarcin, triethylarcin, triphenylarcin, triphenylarcin oxide, ethylenebis (diphenylarcin), tris (dimethyl Amino) arsine, or an arsenic containing compound having the formula As (OR) ₃ , wherein R is -CH ₃ or -C ₂ H ₅ . In some embodiments, the dopant source is As ₂ O ₃ and / or As ₂ O ₅ . In some embodiments, the dielectric material of the second film layer is SiO ₂ .

Some methods disclosed herein may further include forming a fourth conformal film layer containing a dielectric material, and some methods disclosed herein comprise a dopant comprising a dopant on a material adhered to the surface of the substrate. It may further comprise forming a fifth conformal film layer comprising a source, with a substantial portion of the fourth film layer being interposed between the third film layer and the fifth film layer. In certain such embodiments, the methods may further comprise driving a portion of the dopant from the fifth film layer into the substrate to form a conformal doping profile in the substrate. In some such embodiments, forming the fifth film layer includes introducing a dopant precursor into the reaction chamber, depositing the dopant precursor on a material that is adhered to the surface under conditions such that the dopant precursor forms an adsorption limiting layer. Adsorbing, and reacting the adsorbed dopant precursor to form a dopant source. Some methods disclosed herein may further include forming a sixth film layer that is a capping layer, wherein the capping layer is the outermost film layer of the first to sixth film layers relative to the substrate. In certain such embodiments, the capping layer comprises SiO ₂ .

In some of the methods disclosed herein, forming the second film layer comprises introducing a reactant into the reaction chamber, introducing a dielectric precursor into the reaction chamber, under conditions under which the dielectric precursor forms an adsorption limiting layer. Adsorbing the dielectric precursor on a material adhered to the substrate, and reacting the adsorbed dielectric precursor with the reactant to form the dielectric material. In some of the methods disclosed herein, forming the fourth film layer includes introducing a reactant into the reaction chamber, introducing a dielectric precursor into the reaction chamber, under conditions under which the dielectric precursor forms an adsorption limiting layer. Adsorbing the dielectric precursor on a material adhered to the substrate, and reacting the adsorbed dielectric precursor with the reactant to form the dielectric material. In some such embodiments, the reactant may be an oxidant, such as elemental oxygen, nitrous oxide, water, alkyl alcohol, carbon monoxide or carbon dioxide. In some embodiments, reacting the adsorbed dielectric precursor with the reactant to form the dielectric material may include activation using plasma. In some embodiments, the dielectric precursor is an alkylamino silane (SiH _x (NR ₂ ) _4-x ), wherein x = 1-3 and R comprises an alkyl group, and halosilane (SiH _x Y _{4 -x} ), wherein x = 1-3, and Y includes Cl, Br, and I. In some embodiments, the dielectric precursor may be BTBAS.

In some embodiments, each of the first to fifth film layers is a monolayer. In some embodiments, the average thickness of each of the first to fifth film layers is between 0.1 and 2 angstroms. In some embodiments, at least one of the three-dimensional features has a width of about 40 nanometers or less. In some embodiments, the average concentration of dopant on the first to fifth film layers is between about 0.01 and 10 wt%.

A method of doping a patterned substrate in a reaction chamber, comprising directly forming an alloy comprising a dielectric and a dopant source on the surface of a patterned substrate having three-dimensional features or on a material adhered to the surface. Wherein the dopant source comprises a dopant, forming the alloy, and driving a portion of the dopant from the alloy into the substrate to form a conformal doping profile on the substrate. Also disclosed herein. In some such embodiments, forming the alloy includes introducing a dopant precursor and a dielectric precursor into the reaction chamber, adhering onto or to the substrate under conditions where the dopant precursor and the dielectric precursor form an adsorption limiting layer. Co-adsorbing the dopant precursor and the dielectric precursor onto the material, and reacting the adsorbed dopant precursor and the adsorbed dielectric precursor to form the dopant source and the alloy having the dopant source. In some embodiments, forming the alloy may include introducing a third reactant into the reaction chamber and reacting the third reactant with the adsorbed dielectric precursor. In some embodiments, the third reactant may be an oxidant. In some embodiments, reacting the third reactant with the adsorbed dielectric precursor may further comprise activation with plasma.

Also disclosed herein are devices for doping a substrate. In some embodiments, the apparatus includes a reaction chamber, a substrate holder in the reaction chamber, one or more dopant precursor gas inlets connected to the reaction chamber, one or more vacuum pumps, the reaction chamber to the one or more vacuum pumps. It may also include a controller comprising one or more gas outlets to connect, and a machine readable code. In some such embodiments, the code includes instructions for determining that a substrate is present on the substrate holder, directing the dopant precursor into the reaction chamber such that the dopant precursor is adsorbed on the substrate or on a material adhered to the substrate. A film layer containing instructions for operating one or more dopant precursor gas inlets, instructions for operating one or more gas outlets to evacuate the nonadsorbed dopant precursor from the reaction chamber, and a dopant source May comprise instructions for reacting the adsorbed dopant precursor to form a. In some embodiments, the devices disclosed herein may further include an RF generator configured to excite the plasma in the reaction chamber. In certain such embodiments, the instructions for reacting the adsorbed dopant precursor may include instructions for operating the RF generator to excite a plasma that causes the adsorbed dopant precursor to react. In some embodiments, the machine readable code including the controller may further include instructions for driving a portion of the dopant source from the film into the substrate. In some embodiments, the devices disclosed herein may further include a heater configured to heat the film containing the dopant source. In certain such embodiments, the instructions for driving the portion of the dopant into the substrate may include instructions for heating the film to operate the heater to cause thermally mediated diffusion of the dopant from the film into the substrate. It may be.

1 is a schematic illustration of a sequence of operations that may be used to form one or more dopant sources containing conformal film layers via an ALD process.
2 is a schematic illustration of a sequence of operations that may be used to form one or more dielectric conformal film layers through a CFD process.
3 is a schematic illustration of a conformal film stack having interposed ALD B ₂ O ₃ dopant layers and CFD oxide layers, all having a capping layer, all deposited on the surface of the underlying semiconductor substrate.
4 (a) is a schematic illustration of an exemplary A + B stack, with part “B” of the stack serving as a capping layer.
4 (b) is a schematic illustration of a conformal film stack having ALD dopant layers and CFD dielectric layers interposed as alternating nanolaminates and also having a capping layer.
4 (c) is a schematic illustration of a conformal film stack having ALD dopant layers and CFD dielectric layers interposed as alternating monolayers and also having a capping layer.
4 (d) is a schematic illustration of a conformal film stack having alloy layers of a dopant source and a dielectric and also having a capping layer.
FIG. 5 is a graph comparing growth rates on CFD oxide with growth rates on native oxide and existing B ₂ O ₃ monolayers for 50 cycles of deposition.
6 is a schematic illustration of a sequence of operations that may be used to deposit a stack of conformal films and dop an underlying semiconductor substrate.
FIG. 7 is a graph comparing in-wafer (WiW) thickness non-uniformity (NU) between several B ₂ O ₃ stacks deposited using several different methods.
FIG. 8 is the results of 13 point Fourier transform infrared spectroscopy (FTIR) measurements of B ₂ O ₃ films formed using ALD versus CFD, taken at the center and edge of the wafer.
9 shows plots of boron concentration versus substrate depth measured at the substrate center and edge after doping with the conformal stack.
10 also displays plots of boron concentration versus substrate depth measured at the substrate center and edge after doping with a different conformal stack.
FIG. 11A is a schematic illustration of a sequence of operations for doping a lower semiconductor substrate using a conformal film having an arsenic silicate glass (ASG) based dopant layer.
11 (b) is a schematic plot of the anneal temperature profile.

Solid-state diffusion of p-type and n-type dopants from conformally deposited films is presented herein as an alternative to doping substrates using incumbent ion implantation techniques. . Doping through conformally deposited films is, for example, when doping the ultra-shallow junction regions and / or source / drain extension regions of 3D structures, such as 3D transistors (such as pMOS FinFETs). It may be superior to ion implantation. In general, doping through conformally deposited films is useful for doping many high aspect ratio device structures, and conventional ion implantation or directional doping methods may be appropriately used in many scenarios where it is inadequate. Such conformal doping may employ the use of deposition techniques such as atomic layer deposition (ALD). In some contexts, the formation of ALD, called conformal film deposition (CFD), is employed. In particular, when a thin layer of n-doped or p-doped conformal film is formed via ALD on the vertical surfaces of the device structures, conformal doping of these vertical surfaces may be achieved. Such conformal doping has been observed in some cases to increase the current density in 3D devices by 10% to 25% due to the reduced series resistance. See VLSI 2011 by Yamashita et al. In addition to conventional silicon-based microelectronics, other applications of ALD doping include microelectronics and optoelectronics based on III-V semiconductors such as GaAs, and II-VI semiconductors such as HgCdTe, and photovoltaic, flat panel display and electrochromic technologies. It includes.

Doping through the conformally deposited films involves the formation of one or more dopant source conformal film layers containing a dopant source—eg, on the lateral and vertical surfaces of the high aspect ratio device structure. In some embodiments, the dopant source layers may be interposed with silicon oxide conformal film layers, which layers are formed by growing them sequentially over one another, in turn, alternately. If there is a sufficient number of dopant source layers, the dopant may be transferred into the substrate from the dopant source conformal film layers by driving a portion of the dopant into the substrate, such as the underlying silicon substrate. The dopant source driven into the substrate and the dopant source in the conformal film layer (s) may be just the same, or the dopant driven into the substrate may be a species chemically related to the dopant source, for example Note that the dopant source may be a phosphorus containing compound, and the dopant itself may be a phosphorus atom in the phosphorus containing compound.

In some embodiments, it may be possible to use a chemical vapor deposition (CVD) process to deposit dopant source layers. In some embodiments, dopant source layers formed via CVD may be interposed with conformal silicon oxide layers. However, CVD formed dopant source layers thermally grown by flowing B ₂ H ₆ (or PH ₃ or AsH ₃ (arcin) for n-type doping) may, in some cases, be deposited as ( It has been found that the in-wafer uniformity may be less than optimal in terms of both the incorporation of the dopant into the as-deposited source layer and the transfer of the dopant into the underlying substrate. In addition, an array of 3D structures on the same die with different geometric constraints may exhibit pattern loading. That is, features with different geometrical features (eg, different aspect ratios) may have different doping profiles when CVD-based dopant source layers are used.

In some embodiments, atomic layer deposition (ALD) may be used to deposit a dopant source containing conformal film layers. In some embodiments, these layers may be interposed with conformal oxide layers. In some embodiments, oxide or oxyhydride film layers deposited via ALD may be used as dopant source containing conformal film layers. In certain such embodiments, the dopant precursor used to form these layers may be an organometallic precursor. In certain such embodiments, trimethyl borate (TMB) may be used as the ALD dopant precursor in the process of forming boron-containing conformal film layers interposed with oxide layers to form an oxide stack. In other embodiments, arsenic (As) doped silicate glass (ASG) or arsenic oxide (eg, As ₂ O ₃ , As ₂ O ₅ ), or phosphorus (P) doped silicate glass (PSG) or phosphorus oxide ( For example, P ₂ O ₅ ) may serve as a dopant source in the conformal film layer formed through ALD. As described below, plasma activation is used after adsorption of the dopant precursor to produce conformal films containing B ₂ O ₃ , or P ₂ O ₅ or PSG or ASG, or As ₂ O ₃ , or As ₂ O ₅ . It may be. Depending on the embodiment, certain stack engineering concepts may be used to adjust boron, phosphorus, or arsenic concentrations in these films. Of course, other dopant precursors may also be employed in ALD processes to produce dopant source conformal film layers, and, depending on the dopant precursor, a conformal film containing gallium, phosphorus and their respective oxides, oxyhydrides, and the like. Layers may be made.

The dopant source containing conformal film layer is referred to as a “source” layer because it provides a source of dopant species (eg, dopant atoms such as boron, phosphorus, gallium and / or arsenic). The dopant source layer serves as a source of dopant for doping the bottom (or top) structure in the device. After the dopant source layer is formed (or during its formation), the dopant species are driven or otherwise incorporated into adjacent structures in the device being manufactured. In some embodiments, the dopant species are driven by an annealing operation during or after forming the dopant source conformal film layer. As mentioned above, the substantial conformal nature of the dopant source layers deposited using ALD allows for effective doping of non-conventional device structures, including structures that require doping, for example in 3D. In some embodiments, the dielectric layer serves as a base source layer into which the dopant layer is incorporated. For example, doped silicon oxide may be used as the diffusion species for boron, phosphorus, arsenic, gallium, and the like. In some embodiments, boron doped silicate glass (BSG), phosphorus doped silicate glass (PSG), boron-phosphorus doped silicate glass (BPSG), or arsenic doped silicate glass (ASG) or the like may be used.

Conformal films formed using ALD and comprising a dopant source for diffusion into the underlying substrate may have several benefits and advantages. First, in some embodiments, often highly conformal films may be formed by ALD processes at quite low temperatures. The high degree of conformality for these films is beneficial because removal is overetched (or underetched / in the range where the film layers used to dope the underlying layers may be sacrificial and these films are non-conformal). Undercutting), which presents a significant challenge. ALD processes may also provide conformal films with extremely well controlled doping concentrations. In some embodiments, this is achieved by using ALD to sandwich the doped layers with undoped oxide layers and by varying the ratio of doped layer to undoped layer to precisely control the overall stack concentration of the dopant source species. May be That is, the doping level can be tightly controlled by controlling the frequency of the doping layer stacking into the undoped layers and also the conditions of the doping cycle that affect the concentration of the dopant in the dopant source containing layers. For example, in some embodiments, a dopant source with significant steric hindrance may provide a mechanism for controlling the properties of the deposited doped layers.

In addition, conformal films deposited via ALD may have other applications besides being used as a dopant source for the underlying substrate. Other applications for conformal oxide films deposited via ALD include dielectric liners for interlayer dielectric (ILD) applications, intermetal dielectric (IMD) applications, pre-metal dielectric (PMD) applications, and through-silicon via (TSV). Applications, but are not limited to stacked RAM fabrication applications in ReRAM (resistive RAM) applications and / or DRAM.

Other uses for ALD glasses include, but are not limited to, passivation, reflow, hermetic sealing, capacitor dielectrics, and the like.

Further, in some embodiments, the doped films are as etchable layers such as high-etch-rate sacrificial oxide used in various stages in integrated circuit IC fabrication, such as during patterning stages of integrated circuit (IC) fabrication. You can also find applications. In certain such embodiments, the etchable layer for use in IC fabrication may be a glass layer having a tunable wet etch rate where the etch rate is tunable by doping level. That is, the doping level is selected to provide a predefined etch rate. Such doped films may be conformal films deposited via ALD. In certain embodiments, the etchable layer is a silicate glass layer comprising a dopant, such as arsenic, phosphorous, boron, germanium, or a combination thereof. In particular, boron-containing silicate glass (BSG) deposited via ALD has a much higher etch rate than undoped CFD oxide or thermal oxide, and the etch rate of BSG appears to be tunable to boron concentration. Thus, doped conformal films deposited via ALD may simply be used as conformal sacrificial layers or spacers in front-end-of-line (FEOL) structures. Faster etch rates and selectivity in the underlying layer are key features in selecting etch rate selective ALD conformal films.

Formation of Dopant Source Containing Conformal Film Layers via ALD

The concept of an ALD "cycle" is important in relation to the various parts of the following description and in particular the discussion of the deposition of multiple conformal film layers over multiple "cycles" of ALD. In general, the basic set of steps that result in the deposition of at least a partial conformal film layer on the substrate surface may be referred to as a single ALD “cycle”. Typically, the ALD cycle will only include the steps necessary to deliver and adsorb each reactant to the substrate surface and then react these adsorbed reactants to form a partial layer of the film. Of course, the cycle may also include certain minor steps that sweep one of the reactants or by-products out of the reaction chamber and / or apply treatment to the deposited conformal film layer (or partial film layer). In general, a cycle represents a single iteration of a unique set of operations that results in the deposition of a full or partial conformal film layer.

1 shows a schematic diagram illustrating a general sequence of operations from left to right over time along a horizontal axis that may be used to form one or more dopant source containing conformal film layers through an ALD process. This schematic diagram simply illustrates the baseline sequence of steps, in addition to providing ALD-forming conformal film layers, as well as other minor operations depending on the embodiment and depending on the nature of the dopant source and the corresponding dopant precursor species. Note, that substitutions, and / or may be intervened.

Referring to FIG. 1, in operation (i), a suitable dopant precursor is introduced into the reaction chamber under conditions where the dopant precursor may form an adsorption limiting layer (given a sufficient duration of time and a sufficient amount of precursor). And are adsorbed onto the substrate surface (or on the material adhered to the substrate). The duration of operation (i) is selected to allow the precursor to adsorb on the substrate surface in an amount sufficient to support one cycle of ALD film growth. In some embodiments, the precursor saturates the substrate surface forming the adsorption limiting layer described above. In other embodiments, the layer is formed by only partially saturating the substrate surface with the precursor (eg, due to insufficient time or total amount of precursor for saturation). In certain embodiments, the dopant precursor is provided to the reaction chamber by flowing the dopant precursor gas along with the carrier gas. Inert gases such as argon may serve this purpose. Following adsorption of the dopant precursor, in operation (ii), the remaining non-adsorbed dopant precursor still present in the reaction chamber is for example pumped down the reaction chamber to base pressure (referred to as "pump to base"), and And / or optionally, removed by purging the reaction chamber with an inert gas (called an "inert purge") to flush out any remaining nonadsorbed dopant precursor. Thereafter, reaction of the adsorbed dopant precursor occurs in operation (iii), where the reaction proceeds through plasma activation, using inert plasma activation and / or oxidative plasma activation to form a layer of the dopant source containing conformal film. . Finally, in operation (iv), a pump and / or inert purge to another base is performed to remove any reaction byproducts. Operations (i) through (iv) may be referred to as ALD cycles and provide a single cycle of dopant ALD film deposition, as collectively described above. 1 also schematically illustrates that this ALD cycle may be repeated one or more times to form multiple dopant source containing conformal film layers (see operation (v)). Thus, in some embodiments, operations (i) to (iv) are consistently repeated at least once, or at least two times, or at least three times, or at least five times, or at least seven times, or at least ten consecutive times do. Dopant ALD membranes can be between 0.1 kPa and 2.5 kPa per ALD cycle, or between 0.2 kPa and 2.0 kPa per ALD cycle, or between 0.3 kPa and 1.8 kPa per ALD cycle, or 0.5 kPa per ALD cycle. Between or about 1.5 ms, or between 0.1 and 1.5 ms per ALD cycle, or between 0.2 and 1.0 ms per ALD cycle, or between 0.3 and 1.0 ms per ALD cycle, or It may be deposited at rates of or between 0.5 Hz and 1.0 Hz per ALD cycle.

Several dopant precursors may be used with a sequence of previous operations to form dopant source containing conformal film layers. In some embodiments, a suitable dopant source may be an elemental arsenic or arsenic doped silicate glass (ASG), or an arsenic compound such as As ₂ O ₃ and / or As ₂ O ₅ . In other embodiments, suitable dopant sources may be boron compounds such as elemental boron or B ₂ O _3, and suitable dopant precursors may be boron compounds such as alkyl borates. For example, trimethyl borate (TMB) (shown below) is an alkyl borate that acts particularly well as a dopant precursor for forming dopant source B ₂ O ₃ , but other dopant precursors also form various boron-based dopant sources. May be suitable. Also, as described below, dopant sources based on elements other than boron and arsenic, such as, for example, dopant sources based on gallium or phosphorus, may also be suitable.

As indicated above, plasma activation in operation (iii) may involve the exposure of the adsorbed dopant precursor to either an oxidative plasma or an inert plasma. The oxidative plasma may be formed from one or more oxidants such as O ₂ , N ₂ O, or CO ₂ and optionally may include one or more diluents such as Ar, N ₂ , or He. In one embodiment, the oxidative plasma is formed from O ₂ and Ar. Suitable inert plasma may be formed from one or more inert gases such as He or Ar. Note that inert plasma activation may require a longer duration of RF radiation, in some embodiments longer than 10 seconds, to produce a plasma that forms sufficient reactivity in the adsorbed dopant precursor. Without being limited to a particular theory, the longer purge duration may be due to the longer lifetime associated with He (or other inactive) metastable species.

Further, in some embodiments, pumping to the base refers to pumping the reaction chamber to base pressure by directly exposing the reaction chamber to one or more vacuum pumps. In some embodiments, the base pressure may typically be only a few milliTorr (eg, between about 1-20 mTorr). Further, as described above, the pumping step to the base may or may not be performed by an inert purge, so that the carrier gas may flow when one or more valves open the conduction path to the vacuum pump or It may not.

The atomic layer deposition that forms the dopant source containing conformal film layer may be performed at various temperatures. In some embodiments, a suitable temperature in the reaction chamber is between 25 ° C. and 450 ° C., or between 50 ° C. and 300 ° C., or between 20 ° C. and 400 ° C., or between 200 ° C. and 400 ° C. It may be in the range of about or about, or between 100 or 350 degreeC. However, unlike a thermal CVD process involving B ₂ H ₆ , for example, the introduction of only TMB (or other dopant precursor) does not form a layer of B ₂ O ₃ -instead of after surface adsorption of the dopant precursor. Note that the plasma activation step is followed (as described). Atomic layer deposition forming dopant source containing conformal film layers may be performed at various chamber pressures. In some embodiments, a suitable pressure in the reaction chamber is between or about 10 mTorr and 10 Torr, or between 20 mTorr and 8 Torr or about 50 mTorr and 5 Torr or about, or 100 mTorr and 2 Torr. It may be in the range or the like. Various RF power levels may be employed to generate the plasma used in operation (iii). In some embodiments, suitable RF power is between or about 100 W and 10 kW, or between or about 200 W and 6 kW, or between or about 500 W and 3 kW, or between 1 and 2 kW or It may be in that range. Various dopant precursor flow rates may be employed in operation (i). In some embodiments, a suitable flow rate may be between 0.1 mL / min and 10 mL / min, or between 0.5 mL / min and 5 mL / min, or between 1 mL / min and 3 mL / min, or It may be in that range. Different gas flow rates may be used in various operations. In some embodiments, the general gas flow rate may range between 1 L / min and 20 L / min or so, or between 2 L / min and 10 L / min or so. In optional inert purge steps in operations (ii) and (iv), the burst flow employed is between or about 20 L / min and 100 L / min, or 40 L / min and 60 L / It may range between min or so.

When using TMB as a dopant precursor with O ₂ / Ar plasma activation, it may not be appropriate to have O ₂ and TMB co-flowed in the reaction chamber, due to stability aspects (or possibly other aspects). In these embodiments, the pump and / or inert purge step to the base follows the step of introducing and adsorbing the dopant precursor, and thus, the nonadsorbed dopant precursor prior to starting the flow of O ₂ into the reaction chamber. You can also remove If O ₂ is already flowing (perhaps from a previous ALD cycle), the flow may be stopped before the introduction of TMB into the reaction chamber. In some embodiments, it may be appropriate to use a separate line charge step to establish the flow of O ₂ into the chamber after a post-dose purge. A pump (and selective inert purge) step to the base to remove nonadsorbed TMB before introducing O ₂ is known to prevent gas phase CVD reactions. All other process parameters remain the same without the pump to base (and optional inert purge) steps, and the uniformity of the 100X conformal film (100 layers) formed from 100 ALD cycles is degraded from 0.3% to 4%. Figured out.

The adsorption step of the reaction sequences described above (and other similar related reaction sequences) may form an adsorption limiting layer, i.e., the entire dopant (giving a sufficient duration of time and a sufficient amount of precursor)-i.e. a single Conformal film growth in the ALD cycle occurs under conditions that are limited to the amount of dopant precursor that may be adsorbed on the substrate surface at one time. The self-limiting nature of these reactions results in a deposited film that exhibits an excellent degree of conformity and uniformity with respect to the density of the dopant source contained in the film.

As described above, trimethyl borate is one suitable dopant precursor, but depending on the embodiment, other compounds may also serve as suitable dopant precursors. For example, other suitable boron-based dopant precursors include other alkyl borates such as triethyl borate, triisopropyl borate and tri-n-butyl borate, and trimethylboron, triethylboron, triphenylboron, tri-i-propyl borate , Tri-n-amyl borate, B-tribromoborazine, tris (pentafluorophenyl) borane and other similar boron containing compounds. In addition, dopant sources based on elements other than boron may also be suitable. Examples include dopant sources based on gallium, phosphorous, arsenic, or other elements suitable for doping a semiconductor substrate, such as other valence III and V elements. Certain arsenic based dopant sources, which may include conformal films deposited through an ALD process, include arseno-silicate or arsenic-doped silicate glass (ASG), arsenic oxides (eg, As ₂ O ₃ , As ₂ O _5). ) And arsenic oxyhydrides, but are not limited to these. Arsenic based dopant precursors may include, but are not limited to, alkylarcin, alkoxyarcin, and aminoarcin chemistries, and include, but are not limited to, the following specific compounds: arsine, triethylarsenate, trimethylarcin, Triethylarcin, triphenylarcin, triphenylarcin oxide, ethylenebis (diphenylarcin), tris (dimethylamino) arcin, and As (OR) _3, where R is -CH ₃ or -C ₂ H ₅ or other alkyl groups (including saturated or unsaturated alkyl groups)] and other similar arsenic containing compounds, but are not limited to these. Certain phosphorus based dopant sources that may include a conformal film deposited through an ALD process may include, but are not limited to, phosphorus doped silicate glass (PSG). Phosphorous based dopant precursors may include, but are not limited to, triethoxyphosphine oxide, trimethylphosphate, trimethylphosphite, and other similar phosphorus containing compounds. The choice of dopant precursor is typically governed by the ease of integration into existing delivery systems, purity requirements, and overall cost.

In many cases, however, arsenic appears to have distinct advantages over phosphorus for the purpose of forming conformal film layers to dope the underlying substrate. For example, arsenic has a much shorter diffusion length in the silicon lattice compared to phosphorus, and therefore has the potential to dope ultra shallow junctions smaller than 20 nm. Second, the solid solubility of arsenic in silicon is higher than that of phosphorus, indicating that higher concentrations of arsenic dopants can be realized at much narrower junctions. Both of these characteristics are useful in situations that mitigate short channel effects in scaled devices. However, arsenic spin-on-glass (As-SOG) (see Journal of the Korean Physical Society, Vol. 51 (Dec. 2007), pp. S265-S269) is one of the few methods pursued by researchers for ASG. It seems to be one of them, and achieving complete conformity and thickness control by this technology has been a challenge until now and will continue to be a challenge. Thus, forming conformal film layers of arsenic based dopant sources such as arsenic oxide or oxyhydride using the various ALD based methods disclosed herein is a useful and preferred technique for doping certain types of device structures.

Dielectric conformal film layer (eg SiO / SiO) via CFD process ₂ Formation of layers)

Some of the methods disclosed herein for depositing conformal films (and using conformal films to dope an underlying semiconductor substrate) do not contain a dopant source but instead consist of conformal film layers composed of other components. It involves depositing intervening multiple dopant source containing conformal film layers. For example, a dielectric conformal film layer containing SiO / SiO ₂ may be formed in such a way that a deposited dielectric layer is interposed in the dopant source containing conformal film layers as part of the “stack” of the conformal film, as described in detail below, It may be deposited via a conformal film deposition (CFD) process (as shown schematically in FIG. 2). In some embodiments, an ALD dopant layer is deposited at the interface or surface of the underlying semiconductor substrate, followed by undoped protection, where the CFD dielectric layers are interposed with dopant ALD layers every X ALD cycles, which may be a CFD silicon oxide film. It is optionally topping with a "capping" layer. This “stack” configuration is schematically illustrated in FIG. 3 as B ₂ O ₃ serves as the dopant source and serves as the dielectric layers interposed with oxide layers.

Thus, the concept of dielectric CFD "cycle" is important in several parts of the description that follows, and especially with regard to the discussion regarding the deposition of multiple conformal film layers via multiple dopant ALD "cycles". As in the dopant ALD cycle discussed above, the baseline set of operations that result in the deposition of at least a partial conformal film dielectric layer may be referred to as a single dielectric CFD “cycle”. Typically, the dielectric CFD cycle will include only the steps necessary to deliver and adsorb each reactant to the substrate surface and then react these adsorbed reactants to form a partial layer of the dielectric film. Of course, as in the dopant ALD cycle discussed above, the cycle may be such as sweeping one of the reactants or by-products out of the reaction chamber and / or applying treatment to the deposited conformal film layer (or partial film layer). It may also include certain minor steps.

2 may be used to form one or more dielectric conformal film layers through a CFD process and displays a schematic diagram illustrating a general sequence of operations from left to right over time along the horizontal axis. This schematic diagram briefly illustrates the baseline sequence of steps, and note that, depending on the embodiments and the nature of the dielectric film, other minor operations may be added, substituted, and / or intervened in the embodiments described herein. Initially in the sequence, during operation A, a gaseous oxidant is introduced into a reaction chamber containing a semiconductor substrate on which CFD films are to be deposited. Examples of suitable oxidants include elemental oxygen (eg, O ₂ or O ₃ ), nitrous oxide (N ₂ O), water, alkyl alcohols such as isopropanol, carbon monoxide, and carbon dioxide. The oxidant is often provided with an inert gas such as argon or nitrogen.

Next, in operation B, a dielectric precursor is introduced into the reaction chamber. The duration of operation B is selected to allow the precursor to adsorb on the substrate surface in an amount sufficient to support one cycle of film growth. In some embodiments, the precursor saturates the substrate surface. The dielectric precursor is selected from the ability to produce a dielectric of the desired composition. Examples of dielectric compositions include silicon oxide (including silicate glass), silicon nitride, silicon oxynitride and silicon oxycarbide. Examples of suitable dielectric precursors include various alkylamino silanes (SiH _x (NR ₂ ) _4-x ) with x = 1-3 and R includes alkyl groups such as methyl, ethyl, propyl, and butyl in various isomeric configurations. do. Examples of suitable dielectric precursors also include several halosilanes (SiH _x Y _4-x ), where x = 1-3 and Y includes Cl, Br, and I. More specific examples of suitable dielectric precursors include various bis-alkylamino silanes and sterically hindered alkyl silanes. In some embodiments, bis (tert-butylamino) silane (BTBAS) is a suitable dielectric precursor for producing silane oxide.

The oxidant that was introduced into the chamber during operation A may continue to flow during operation B. In certain embodiments, the oxidant continues to flow at the same concentration and at the same rate as during operation A. Upon completion of operation B, the flow of the dielectric precursor into the chamber is terminated and operation C begins as shown. During operation C, the oxidant (and accompanying inert gas, if present) may continue to flow as during operations A and B to purge potentially remaining nonadsorbed dielectric precursor from the reaction chamber. In other embodiments, the flow of oxidant may be shut off and purge may be performed only with an inert gas. In some embodiments (not shown in FIG. 2), removal of the non-adsorbed dielectric precursor from the reaction chamber may be achieved through a pump operation to a base similar to the operation described above.

After removing the nonadsorbed dielectric precursor during operation C, in operation D, the precursor reacts on the substrate surface to form a portion of the dielectric film. In various embodiments, operation D may involve applying a plasma to the adsorbed dielectric precursor to drive the reaction. In some embodiments, the reaction is an oxidation reaction. Some of the oxidant flowing into the reaction chamber may be adsorbed onto the substrate surface along with the dielectric precursor, thus providing a readily available oxidant for plasma mediated surface reactions. In some embodiments, the flow of the oxidant into the reaction chamber may continue through the duration of operation D (optionally) as shown in FIG. 2, and may drive this flowing oxidant to drive oxidation of the adsorbed dielectric precursor. An oxidative plasma may be formed.

Operations A through D collectively provide a single cycle of dielectric CFD film deposition. However, it should be understood that other CFD processes may be used instead of the basic cycle described herein. In the described embodiment, the CFD cycle (consisting of operations A to D) is performed without the introduction of any dopant species or dopant precursors (of course dopants may be introduced). In various embodiments, the dielectric CFD cycle represented by operations A through D is repeated one or more times in succession prior to the introduction of any dopant species or dopant precursor. This is schematically represented by operation E of FIG. Thus, in some embodiments, operations A through D are repeated at least once or at least twice or at least three times, or at least five times, or at least seven times or at least ten times before initiating the dopant ALD cycle as described above. . CFD dielectric films can be between 0.1 kPa and 1.5 kPa per CFD cycle, or between about 0.2 kPa and 1.0 kPa per CFD cycle, or between 0.3 kPa and 1.0 kPa per CFD cycle, or 0.5 kPa per CFD cycle. The film may be deposited at a rate between or about 1.0 Hz. Over each of the one or more CFD cycles, the oxidant may continue to flow in the reaction chamber as shown schematically in FIG. 2.

In certain embodiments, plasma activation involves RF power of any frequency suitable for causing a reaction of the adsorbed dielectric precursor in operation D. In some embodiments, RF power may include independently controlled high frequency RF power and low frequency RF power generated by an RF power source that may be operated independently of each other. Suitable low frequency RF powers may include, but are not limited to, a frequency range between about 200 kHz and 1000 kHz. Suitable high frequency RF powers may include, but are not limited to, frequencies between about 10 MHz and 80 MHz, such as 13.56 MHz. As such, RF power supplies and matching networks may be operated in any suitable power range to form a plasma in the reaction chamber. Examples of suitable plasma generated RF power may include RF power between about 100 W and 3000 W for high frequency plasma (based on per wafer), and RF power in the range between about 100 W and 10,000 W for low frequency plasma. It may, but is not limited to these. One or more RF power supplies may be operated at any suitable duty cycle. Examples of any suitable duty cycle include, but are not limited to, duty cycles between about 5% and 90%. These RF powers and frequencies may also be suitable for use in the plasma activation operation of the dopant ALD cycle described above.

Reaction chamber pressures suitable for carrying out operations A to D generally include, but are not limited to, pressures between 0.5 Torr and 5 Torr, or preferably between 2 Torr and 4 Torr. . It has been found that pressures up to about 10 Torr (or up to about 9 Torr) work well for certain plasma pretreatment (of the lower substrate) prior to exposure to the dopant.

Stack Engineering: Capping Floor

In some embodiments, the dopant source containing conformal film may be hygroscopic with the ability to adsorb sufficient ambient humidity to affect the physical properties and functions observed. For example, a conformal membrane containing B ₂ O ₃ , As ₂ O ₃ , As ₂ O ₅ , P ₂ O ₅ , PSG, or ASG, deposited through ALD, may be hygroscopic and may be moisturized such as volatile boric acid. Hydroxide may form upon exposure to water. Volatilization of boric acid changes the ability of these films to dope the underlying substrate. Thus, it is often advantageous to protect the dopant source containing conformal film layers from exposure to moisture.

Thus, prevention or minimization of moisture exposure is deposited over the lower dopant source layer or stack of layers, essentially acting as a physical barrier between these layers and any ambient moisture present in the reaction chamber, as illustrated in FIG. 3. It may be achieved by employing a capping layer that is. 4 (a) shows an exemplary A + B stack and portion “B” of the stack serves as a capping layer. In some embodiments, the capping layer may include one or more oxides, such as SiO and SiO ₂ , formed from a dielectric precursor deposited and reacted in the CFD process as described above. In some embodiments, the dielectric precursor may be bis (tert-butylamino) silane (BTBAS). In other embodiments, the capping layer may comprise one or more nitride or oxynitride species. In some situations, it has been found that without a capping layer, it is not possible to detect the presence of B ₂ O ₃ by thickness metrology.

In some embodiments, the capping layer is engineered by controlling the capping layer chemistry during the deposition process to adjust the diffusion characteristics of the dopant towards the underlying semiconductor substrate. Without wishing to be bound by any particular theory, by depositing a suitable capping layer, dediffusion of the dopant source (ie, diffusion of the dopant away from the membrane-substrate interface) may be blocked by the capping layer, thereby preventing infinite storage of the dopant in the stack. It is assumed that the simulation may be effective.

Stack Engineering: Deposition of multiple dopant source film layers through multiple ALD cycles interposed with other film layers such as dielectric CFD layers

It is possible to deposit multiple dopant source containing conformal film layers sequentially through ALD processes, and in some embodiments and in some types of dopant precursors, this may be the preferred approach. In some cases, however, it has been found that depositing dopant film layers sequentially using sequential ALD cycles tends to be somewhat self-inhibiting. In particular, it has been found that in TMBs that serve as dopant precursors, magnetic inhibition limits TMB nucleation on B ₂ O ₃ monolayers to less than 0.2 kPa per ALD cycle. Also, this growth rate is further lowered after a certain number of dopant layers have been deposited. Indeed, in one embodiment, it was found that the total thickness of the B ₂ O ₃ film using sequential ALD cycles did not vary significantly between the 50th and 100th ALD cycles. As a result, in some dopant sources and their corresponding dopant precursors, the dopant concentration in the deposited conformal film cannot be increased simply by increasing the dopant source concentration in the film or by increasing the film thickness. For example, referring to FIG. 4 (a), the total amount of B ₂ O ₃ available cannot simply be increased by increasing the thickness of the “A” layer in the “A + B” stack described above. This is in contrast to boron-containing layers deposited using a CVD process, where the boron concentration is increased by increasing the duration of the CVD process and / or increasing the flow of dopant precursor to increase the thickness of the deposited boron-containing layer. It can simply be adjusted.

However, in some cases the dopant source containing film layer may experience faster growth when formed through ALD on top of some non-dopant containing film layers, for example, a non-dopant dielectric layer, or more particularly a non-dopant oxide layer. I found out. In addition, when the above-described dielectric layer (eg, oxide layer) is deposited through a CFD process, the dopant source layer formed through ALD thereon is much more than on the natural dielectric layer (eg, natural oxide layer). I found that it might show faster growth. For example, it was found that B ₂ O ₃ dopant source layers formed from TMB precursor deposited via ALD on top of a CFD deposited silicon oxide layer show a growth rate greater than 1 Hz per ALD cycle, as shown in FIG. 5. . Without being limited to a particular theory, this may be due to the greater number of available hydroxyl surface sites on which TMB may adhere onto the coalesced CFD oxide layer. In some embodiments, the “initiation” layer of CFD oxide is deposited on top of the substrate surface prior to depositing the first ALD dopant layer to promote growth of the ALD dopant layer and increase the boron concentration closest to the dopant source-substrate interface. It may be formed. 5 also shows the growth rate for native oxide and conventional B ₂ O ₃ monolayers over CFD oxide over 50 deposition cycles at a temperature of 400 ° C. using an oxidative plasma (O ₂ ) to activate the adsorbed TMB precursor. Compare the growth rate. As shown in the figure, the dopant ALD layers grow at a rate of 0.3 kW per ALD cycle (up to 1 kW per ALD cycle on CFD oxide) on top of the native silicon oxide layer but ALD on top of a conventional B ₂ O ₃ monolayer It was found that only 0.21 ms per cycle grew. Thus, non-CFD deposited silicon oxide or hydroxyl-terminal containing oxide layers may also promote or increase nucleation rates of ALD dopant layer growth, but at least in some cases to the same extent.

Because of this, by nucleating TMB ALD layers on CFD silicon oxide thin layers, the effective boron concentration can be increased for the same overall stack thickness. By varying the spacing and / or frequency of these CFD oxide layers throughout the stack, it is possible to fine tune the boron concentration. 4 (b) and 4 (c) show different stacking strategies, such as employing nanolaminates and monolayers, respectively. Single layer laminates are made by alternating single cycles of CFD silicon oxide deposition and ALD dopant layer deposition. Nanolaminates are prepared by alternating two or more cycles of ALD dopant layer deposition and two or more cycles of CFD silicon oxide layer deposition. In some cases, more than two sequential cycles of dopant layer and / or oxide layer deposition may be employed in forming nanolaminates. For example, the repeating sequence of deposition cycles used to construct the nanolaminate is 2: 2 (ie 2 oxide cycles after 2 dopant cycles, this sequence is repeated one or more times), or 2: 3, or 2 : 5, or 3: 2, 5: 5, or the like. Thus, by adjusting the deposition sequence and the ratio of dopant cycles to oxide cycles, the dopant concentration (eg, boron concentration) for a given stack thickness (eg, about 100 kPa-200 kPa) may be adjusted and controlled.

Thus, FIG. 6 displays a schematic diagram illustrating a sequence of operations that may be used to deposit a stack of conformal films and dope an underlying semiconductor substrate. Again, the sequence of actions progresses from left to right over time along the horizontal axis. Again, like FIG. 1 and FIG. 2, this schematic diagram merely shows a baseline sequence of steps, and depending on the embodiment and various considerations, other additional operations may be added to, substituted for, or substituted for, and / or illustrated. It is also noted that it may be intervened. Such considerations may include, for example, the nature of the design of the conformal stack, the selection of the dopant source and the corresponding selection of the corresponding dopant precursor species, and also the selected composition of the dielectric layer.

Referring to FIG. 6, in this particular embodiment, one or more dielectric layers are first used on the surface of the underlying semiconductor substrate, using a dielectric CFD cycle schematically illustrated in FIG. 6 by operations A, B, C, and D. FIG. Deposited directly on (or on a material adhered to the substrate). The operations A, B, C, and D in FIG. 6 correspond to the A, B, C, and D operations schematically illustrated in FIG. 2, and the foregoing discussion regarding those operations also applies here. As in FIG. 2, operations A through D represent one dielectric CFD cycle of film growth, and as also in FIG. 2, operation E in FIG. 6 indicates that additional dielectric CFD cycles follow the initial cycle to further conformal dielectric film layer. It is shown schematically that they may be formed. As such, the baseline process schematically illustrated in FIG. 6 includes and begins with the baseline dielectric CFD process illustrated in FIG. 2.

Following deposition of one or more dielectric film layers in operations A through E, in operation F the reaction chamber is pump-to-base (as described above) and optionally (as described above) Purge with inert gas. Operation F serves to remove any remaining dielectric precursors, oxidized species, or any other remaining species from preceding dielectric CFD cycles from the reaction chamber. Of course, in some embodiments, the removal operation F may not be necessary because, for example, there is no significant amount of remaining species in the reaction chamber after conclusion of the dielectric CFD cycles, or simply the remaining species. This is because they do not adversely affect the dopant ALD cycles to be followed.

After the first set of dielectric film layers are deposited and the reaction chamber is prep (if necessary), the first dopant ALD cycle is performed in operations G, H, I, and J. Operations G-J correspond to operations (i)-(iv) of FIG. 1, and the discussion relating to operations (i)-(iv) of FIG. 1 also applies here to operations G-J. As described above with respect to operations (i) to (iv), in operation G, under conditions where the dopant precursor may form an adsorption limiting layer, a suitable dopant precursor is introduced into the reaction chamber and on the substrate surface (or Adsorbed onto the material adhered to the substrate). After adsorption of the dopant precursor, in operation H, any remaining nonadsorbed dopant precursor still present in the reaction chamber is removed through a pump to the base and an optional inert purge. The reaction of the adsorbed dopant precursor then takes place in operation I, where the reaction proceeds through plasma activation using inert plasma activation and / or oxidative plasma activation to form a layer of the dopant source containing conformal film. Finally, in operation J, a pump and / or inert purge to another base is performed to remove any reaction byproducts. As described above with respect to operations (i) through (iv), operations G through J collectively represent a single dopant ALD cycle providing a single dopant ALD film layer. As schematically represented by operation K of FIG. 6, the ALD cycle represented by operations G through J may be repeated one or more times to form multiple dopant source containing conformal film layers.

At this point in the baseline conformal stack building process schematically illustrated in FIG. 6, the lower semiconductor substrate is a set of dielectric CFD layers deposited thereon and another set of dielectric CFD layers deposited thereon. Dopant ALD layers. The set of dielectric CFD layers and the set of dopant ALD layers may each have the same number of layers, as described above, or in some embodiments, have a different number of layers. In either case, this set of dopant ALD layers may constitute an extent of the dopant sources present in the conformal stack. In other words, in some embodiments, additional dopant ALD layers may not be deposited. However, other embodiments may have multiple sets of dopant ALD layers. Thus, for example, in operation L, the sequence of operations A through K may be repeated one or more times, resulting in one or more additional sets of dielectric CFD layers that alternate with one or more additional sets of dopant ALD layers. The actual number of times operations A to K are repeated depends on the desired total thickness of the conformal film stack, the desired thickness of the dielectric and dopant layers, and the amount / concentration of the dopant intended to be included in the stack. Also, strictly following the schematic process of FIG. 6, the second set of dielectric CFD layers will have the same number of layers as the first set of dielectric CFD layers, and similarly the second set of dopant ALD layers Note that it will have the same number of layers as one set of dopant ALD layers. However, it will be appreciated that the number of layers in each set of continuous dielectric CFD and dopant ALD layers may vary, and in some embodiments, varying the number of layers in this manner may be the preferred approach.

After deposition of the dielectric and dopant conformal film layers is complete, an optional capping layer may be applied in operation M of FIG. 6. Potential capping layers are described above. As shown in FIG. 6, in some embodiments, the capping layer may be formed by repeating the dielectric CFD cycle of operations A-D one or more times. Note, however, that the capping layer does not necessarily need to be formed by employing the operations A-D. In some embodiments, the capping layer can be formed by employing a CFD oxide chemistry that is completely different than that employed in operations A through D; For example, the precursors and / or plasma may be different. In some embodiments, the CFD deposited capping layer may employ one or more CFD nitride or oxynitride film layers.

After the capping layer was deposited, the dopant source containing conformal stack according to FIG. 6 was built / deposited. In some embodiments, for example, the conformal stack may be similar to the conformal stack shown schematically in FIG. 3 where the dopant source is B ₂ O ₃ and the dielectric is one or more oxides. Finally, with the conformal stack fully assembled in operation N, the lower substrate may be doped by driving the dopant from the conformal stack into the lower semiconductor substrate. In some embodiments, this may be accomplished by a thermal mediated diffusion process, for example, as detailed below.

Finally, another stack strategy that may be used to combine CFD dielectric (eg, silicon oxide) layer deposition and ALD dopant layer deposition is to include “alloy layers” as shown in FIG. 4 (d). It is what constitutes a "stack". For example, the alloy layers may be constructed by having dielectric precursor BTBAS and dopant precursor TMB co-flow and oxidize simultaneously in the reaction chamber. That is, in some embodiments, the alloy layer introduces the dopant precursor and the dielectric precursor into the reaction chamber and deposits both precursors on the substrate (or adhered to the substrate) under conditions in which the dopant precursor and the dielectric precursor form an adsorption limiting layer. And by adsorbing the adsorbed dopant precursor and the adsorbed dielectric precursor to form an alloy layer. In certain such embodiments, the alloy layer is formed by reacting the dopant and dielectric precursors with the adsorbed dielectric precursor with a third reactant, which may be an oxidant introduced into the reaction chamber. In certain such embodiments, the reaction of the adsorbed dielectric precursor with the oxidant may include activation with plasma. In other embodiments, the "alloy layers" shown in FIG. 4 (d) deposit the nanolaminates or monolayers shown in FIGS. 4 (b) and 4 (c), respectively, and (typically higher temperatures). However, in some embodiments it may be formed by annealing the layers) at a temperature below the “drive” temperature for forming “alloy layers”, resulting in a composite homogeneous glass layer.

Dopant driving from the deposited conformal film stack into the underlying substrate

After the conformal film stack is completely deposited, the conformal film stack may be used as a source of dopant species for nearby semiconductor structures. This may be accomplished by driving the dopant from the deposited film into the device structure. In various embodiments, driving is accomplished by a thermal mediated diffusion process such as annealing. In some cases, laser spike annealing may be employed, particularly in the case of employing ultra shallow junctions.

Devices

Any suitable apparatus may be employed to make and use the conformal dopant source containing ALD film layers described herein. For example, US patent application No. The apparatus described in 13 / 224,240 may be used. This patent application is incorporated herein by reference in its entirety for all purposes.

In some embodiments, the system controller (which may include one or more physical or logical controllers) controls all activities of the process tool used in the dopant ALD cycle or the dielectric CFD cycle. The system controller executes system control software on one or more processors. System control software includes timing, mixture of gases, chamber and / or station pressure, chamber and / or station temperature, wafer temperature, target power levels, RF power levels, substrate pedestal, chuck and / or susceptor ( susceptor) position, and instructions for controlling other parameters of a particular ALD / CFD process performed by the process tool. System control software may be configured in any suitable manner. For example, various process tool component subroutines or control objects may be created to control the operation of the process tool components needed to execute the various process tool processes. The system control software may be coded in any suitable computer readable programming language.

In some embodiments, the system control software includes input / output control (IOC) sequencing instructions for controlling the various parameters described above. For example, each phase of an ALD / CFD process may include one or more instructions for execution by the system controller. Instructions for setting process conditions for an ALD / CFD process phase may be included in the corresponding ALD / CFD recipe phase. In some embodiments, ALD / CFD recipe phases may be arranged sequentially so that all instructions for an ALD / CFD process phase are executed concurrently with that process phase.

Other computer software and / or programs may be employed in some embodiments. Examples of programs or sections of programs for this purpose include a substrate positioning program, a process gas control program, a pressure control program, a heater control program and a plasma control program.

The substrate positioning program may include program code for process tool components used to load the substrate onto the pedestal and control the spacing between the substrate and other portions of the process tool.

The process gas control program may include code for controlling gas composition and flow rates and optionally code for flowing gas into one or more process stations prior to deposition to stabilize the pressure in the process station. The pressure control program may include code for controlling the pressure at the process station, for example, by adjusting the throttle valve in the exhaust system of the process station, gas flow into the process station, and the like.

The heater control program may include code for controlling the current to the heating unit used to heat the substrate. Alternatively, the heater control program may control the supply of heat transfer gas (such as helium) to the substrate.

The plasma control program may include code for setting RF power levels applied to process electrodes at one or more process stations.

In some embodiments, there may be a user interface associated with the system controller. The user interface may include a display screen, graphical software displays and / or process conditions of the apparatus, and user input devices such as pointing devices, keyboards, touch screens, microphones, and the like.

In some embodiments, the parameters adjusted by the system controller can relate to process conditions. Non-limiting examples include process gas composition and flow rate, temperature, pressure, plasma conditions (eg, RF bias power levels), pressure, temperature, and the like. These parameters are provided to the user in the form of a recipe, which may be entered using the user interface.

Signals for monitoring the process may be provided by analog and / or digital input connections of the system controller from various process tool sensors. Signals for controlling the process may be output on the analog and digital output connections of the process tool. Non-limiting examples of process tool sensors that may be monitored include mass flow controllers, pressure sensors (eg, manometers), thermocouples, and the like. Appropriately programmed feedback and control algorithms may be used with data from these sensors to maintain process conditions.

The system controller may provide program instructions for implementing the dopant ALD and dielectric CFD cycles and processes described above. Program instructions may control various process parameters, eg, DC power level, RF bias power level, pressure, temperature, and the like. The instructions may control these parameters to operate in situ deposition of the conformal film stacks in accordance with various embodiments described herein.

Thus, a system controller for providing program instructions to an apparatus for doping a substrate includes instructions for detecting the presence of a substrate on a substrate holder in a reaction chamber, on a material on which a dopant precursor is adhered to or adhered to the substrate. Instructions for operating one or more dopant precursor gas inlets to introduce the dopant precursor into the reaction chamber to be adsorbed, instructions for operating one or more outlets to evacuate the nonadsorbed dopant precursor from the reaction chamber, and a dopant source. May have machine readable code that may include instructions for reacting an adsorbed dopant precursor to form a film layer. In some embodiments, the instructions for reacting the adsorbed dopant precursor include instructions for operating the RF generator to excite a plasma that causes the adsorbed dopant precursor to react. In some embodiments, the machine readable code may further include instructions for driving a portion of the dopant source from the film into the substrate, and in certain such embodiments, these “drive” instructions heat the film. The method may further include instructions for operating the heater to cause thermal mediated diffusion of the dopant from the film into the substrate.

The apparatus / process described above may be used with lithographic patterning tools or processes, for example, for the fabrication or manufacture of semiconductor devices, displays, LEDs, photovoltaic panels, and the like. Typically, but not necessarily, these tools / processes will be used or implemented together in a normal production facility. Lithographic patterning of a film typically includes some or all of the following operations, each of which is made possible by a number of possible tools: (1) a spin-on or spray-on tool on a workpiece, ie a substrate; Photoresist coating using; (2) photoresist curing using a hot plate or furnace or UV curing tool; (3) photoresist exposure to visible or UV or x-ray light using a tool such as a wafer stepper; (4) resist development using a tool such as a wet bench to selectively remove resist and pattern it; (5) resist pattern transfer to the underlying film or workpiece by using a dry or plasma assisted etching tool; And (6) resist removal using tools such as RF or microwave plasma resist strippers.

Experiments and Examples

7 compares in-wafer (WiW) thickness non-uniformity (NU) between various B ₂ O ₃ stacks deposited using several different methods. Each stack includes conformal film layers of ALD deposited dopant source B ₂ O ₃ interposed with dielectric CFD layers. The first four B ₂ O ₃ stacks, MOD1 to MOD4, were deposited via CVD processes with varying deposition temperatures and B ₂ H ₆ flow rates, these stacks exhibiting a NU range of ˜8-16%. Two B ₂ O ₃ stacks deposited via ALD, in which the ALD1 process employs an inert plasma and the ALD2 process employs an oxidizing plasma, exhibit a drastically improved thickness NU range of ˜1.3-1.6%. Note that the bars of the “% STDDEV” label shown in the plot of FIG. 7 represent the percent standard deviation in NU measured across multiple wafers.

7 also shows the reduction in WiW NU achievable by boron-containing conformal films formed using ALD, relative to the boron-containing conformal films formed via CVD. In FIG. 8, the results of 13 point Fourier transform infrared spectroscopy (FTIR) measurements of B ₂ O ₃ films formed using ALD versus CFD are shown. The results for the CVD film are plotted on the left graph and the results for the ALD films are plotted on the right. Two curves occupy each graph, one corresponding to the FTIR adsorption spectrum taken at the center of the wafer and the other corresponding to the absorption spectrum taken near the wafer edge. For CVD films, the visible difference in center and edge adsorption curves (particularly near boron-oxygen stretching vibrations) can be seen in FIG. 6 corresponding to 10.3% NU. In contrast, the center and edge adsorption curves taken from the ALD membranes almost overlap, exerting only 2.5% NU, again indicating the improved WiW uniformity achievable using ALD membranes. These improvements are mainly due to the substantially perfect surface saturation potentially achievable through ALD processes, especially at the thickness regime (about 5-10 nm) of interest.

Another method of measuring wafer inhomogeneity (WiW NU) at the film stack thickness is to drive a dopant from the stack into the silicon substrate (eg, by spike annealing) and to measure the resulting sheet resistance (R _s ) of the silicon substrate. will be. Since the amount of dopant available to drive into the substrate is correlated with the stack thickness, uniform sheet resistance exhibits a uniform stack thickness. In addition, the WiW uniformity of the dopant concentration after annealing is clearly an important consideration in device fabrication. Table 1 shows the average sheet resistance of silicon substrates after stack deposition and spike annealing corresponding to boron-containing stacks (CVD-B) deposited using the CVD methodology and boron-containing stacks (ALD-B) deposited using the ALD methodology. (Mean R _s ) and percent standard deviation (% Std Dev). Table 1 shows that within the desired dopant drive values typically employed in semiconductor fabrication, ie 100-5000 ohms / sq, the ALD film exhibits a sharply improved WiW NU over sheet resistance compared to the CVD film. When using ALD, the in-wafer change in average sheet resistance can be achieved below the standard deviation of less than 4%, in contrast to the standard deviation being nearly 195% when using CVD.

Table 2 shows the range in which the various stack strategies described above may be used to adjust and control the dopant concentration. In this particular example, the boron concentration was measured in conformal stacks of approximately 100 kPa thick of ALD B ₂ O ₃ and CFD silicon oxide.

As such, by employing different stack strategies, the dopant concentration may be effectively controlled in regimes over two orders of magnitude. Control of stack strategies and stack engineering, and engineering of capping layer chemicals may also be used to adjust and control the degree of sheet resistance and its corresponding WiW NU. For example, Table 3 shows the effect of controlling the ratio and sequence of ALD dopant source layer to CFD oxide layer deposition. Specifically, Table 3 shows the effect of modulating the layer ratio and sequence on the thickness non-uniformity (Thx NU) and sheet resistance WiW non-uniformity (R _s (drive) NU) of the deposited ALD films. Note that the "2: 2" ratio in the table corresponding to the configuration with the lowest sheet resistance WiW NU means that two SiO ₂ layers and the like were deposited after the two B ₂ O ₃ layers were deposited. Note that these ALD film stacks were deposited at approximately 200 ° C., which resulted in consistently lower R _s NUs compared to ALD film stacks deposited at higher temperatures such as 400 ° C., for example.

Table 4 shows that by changing the capping layer chemistry, WiW NU may be improved from 41% to 21%.

Without being bound by a particular theory, back-diffusion of the dopant source (i.e. diffusion of the dopant away from the membrane-substrate interface) by depositing a suitable capping layer is blocked by the capping layer, thereby allowing the dopant in the stack. It is also assumed that we can effectively simulate the infinite store of. 8 schematically illustrates a conformal film stack having a capping layer deposited over a pattern with alternating ALD dopant source and oxide layers deposited on a silicon substrate.

Dopant uniformity in the underlying substrate after transfer from the conformal film stack via spike anneal is illustrated by the boron concentration profiles shown in FIGS. 9 and 10. Both figures plot the boron concentration in the underlying silicon substrate as a function of the substrate depth measured at both the center and edge of the substrate. 9 shows the results of doping a substrate with a 2: 1 ratio stack of Table 3 (with WiW NU of 13.2%). In the figures, some slight differences between the center and edge dopant concentration profiles as a function of depth are clearly shown. FIG. 10 shows the results of doping a substrate with a 2: 2 ratio stack of Table 3 (WiW NU is 5.3%). As shown in the figure, in the case of a 2: 2 stack, the boron concentration profiles nearly overlap to a depth of approximately 200 mm 3. Both figures show high concentrations of boron on the order of 1E22 atoms / cm ³ near the substrate surface, indicating good transfer from the conformal film stacks. The total boron dose is in the range of 8.4E14 atoms / cm ² to 4E15 atoms / cm ² .

As discussed above, an arsenic or phosphorous dopant source may also be used in the conformal film stack to dope the underlying semiconductor substrate. For example, FIG. 11A is a schematic diagram of a sequence of operations for doping a lower semiconductor substrate using a conformal stack having an arsenic silicate glass (ASG) based dopant layer. In this particular embodiment, an ASG based conformal stack is formed by first depositing an ASG conformal film through an ALD process and then forming an undoped silicate glass (USG) capping layer through a CFD process. Note that, depending on the embodiment, various arsenic or phosphorus conformal films may be combined with various kinds of capping layers, such as, for example, various kinds of capping layers disclosed herein. It is also noted that test vehicles similar to those depicted schematically in FIG. 11A may be employed using phosphorus silicate glass (PSG) or boron silicate glass (BSG).

With the stack in place, thermal annealing is performed to drive the arsenic dopant into the underlying substrate. 11 (b) schematically shows an anneal temperature profile. In some embodiments, the temperature ranges for anneal may be between 800 degrees Celsius and 1100 degrees Celsius, or about, or between about 850 degrees Celsius and 1000 degrees Celsius, or between about 875 degrees Celsius and 925 degrees Celsius. In some embodiments, the anneal temperature and anneal time (ie, the period of time to maintain the anneal temperature) may be selected to achieve a particular target sheet resistance in the underlying substrate. In some embodiments, an anneal temperature of approximately 900 ° C. and an anneal time of less than 5 seconds may be selected to achieve a sheet resistance of approximately 600-2000 ohm / sq. FIG. 11 (b) is schematically ramped up to (and below) the maximum temperature over the duration of the annealing process, for some duration where the temperature defines the temperature profile characteristic of the annealing. Again, in some embodiments, this temperature profile may be selected to derive the desired sheet resistance in the underlying substrate.

After the anneal results in the transfer of the dopant from the conformal stack into the substrate, in some embodiments, the stack is removed through a stripping operation. In some embodiments, as shown in FIG. 11 (a), the stripping operation includes treatment of the conformal stack with dilute hydrofluoric acid (HF).

Other embodiments

Many variations on this baseline process may be realized. Some of these variations have an increase in the amount of dopant available for diffusion into adjacent semiconductor structures as a goal. Other variations are designed to control the flow rate at which the dopant is transferred from the source film into the nearby semiconductor structure. Still other variations control the direction in which dopant species diffuse. Frequently, it is desirable to help diffuse the dopant towards the device structure and away from the opposite side of the film. Certain variations are described in US patent application Ser. No. 13 / 242,084, previously incorporated by reference.

In addition, it is to be understood that the configurations and / or approaches described herein are illustrative in nature and that these particular embodiments or examples are not to be considered in a limiting sense because various modifications 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 parallel in other sequences, in the illustrated sequence, or in some cases may be omitted. Likewise, the order of the processes described above may be changed.

Claims (47)

A method of doping a patterned substrate in a reaction chamber,
Forming a first conformal film layer comprising a dopant source comprising a dopant that is boron, either directly on a surface of a patterned substrate having three-dimensional features or on a material adhered to the surface, the method comprising:
Introducing a dopant precursor, which is an alkyl borate, into the reaction chamber;
Adsorbing the dopant precursor directly on the surface or on a material adhered to the surface under conditions in which the dopant precursor forms an adsorption-limited layer; And
Forming the first conformal film layer, comprising reacting the adsorbed dopant precursor to form the dopant source; And
Driving a portion of the dopant from the first conformal film layer into the substrate to form a conformal doping profile on the substrate. The method of claim 1,
Forming the first conformal film layer comprises atomic layer deposition. 3. The method of claim 2,
Forming the first conformal film layer further comprises removing a nonadsorbed dopant precursor from the reaction chamber. The method of claim 3, wherein
Wherein said removing occurs after said adsorbing but before said reacting. 5. The method of claim 4,
Said removing comprises pumping a non-adsorbed dopant precursor from said reaction chamber by pumping said reaction chamber at a base pressure using a vacuum pump. 5. The method of claim 4,
Wherein said removing comprises purging the reaction chamber with an inert gas. The method of claim 1,
Forming a second conformal film layer containing a dielectric material; And
Forming a third conformal film layer on the material adhered to the surface of the substrate, the third conformal film layer comprising a dopant source comprising a dopant,
Introducing the dopant precursor into the reaction chamber;
Adsorbing the dopant precursor on a material adhered to the surface under conditions in which the dopant precursor forms an adsorption limiting layer; And
Reacting the adsorbed dopant precursor to form the dopant source, wherein a substantial portion of the second conformal film layer is interposed between the first and third conformal film layers. Forming the third conformal film layer; And
Driving a portion of the dopant from the third conformal film layer into the substrate to form a conformal doping profile in the substrate. The method of claim 7, wherein
Adsorbing the dopant precursor when forming the first conformal film layer comprises atomic layer deposition, and adsorbing the dopant precursor when forming the third conformal film layer includes atomic layer deposition Doping the patterned substrate. The method of claim 8,
Forming the first conformal film layer further includes removing the nonadsorbed dopant precursor from the reaction chamber, and forming the third conformal film layer comprises the nonadsorbed dopant from the reaction chamber. Removing the precursor, further comprising the step of doping the patterned substrate. The method of claim 9,
During each formation of the first conformal film layer and the third conformal film layer, the removing step occurs after the adsorbing step but before the reacting step. 11. The method of claim 10,
Said removing comprises pumping a non-adsorbed dopant precursor from said reaction chamber by pumping said reaction chamber at a base pressure using a vacuum pump. The method of claim 9,
Wherein said removing comprises purging the reaction chamber with an inert gas. The method of claim 7, wherein
Reacting the adsorbed dopant precursor to form the first conformal film layer includes activation using plasma, and reacting the adsorbed dopant precursor to form the third conformal film layer And activating with the plasma. The method of claim 13,
Wherein the plasma is an oxidative plasma. The method of claim 13,
Wherein the plasma is an inert plasma. The method of claim 7, wherein
Driving a portion of the dopant from the first conformal film layer and the third conformal film layer comprises a thermal mediated diffusion process. 17. The method of claim 16,
And wherein said thermally mediated diffusion process involves annealing. The method of claim 17,
Wherein said annealing is laser spike annealing. The method of claim 1,
Wherein said alkyl borate is trimethyl borate. The method of claim 1,
Wherein the dopant source is boron oxide or boron oxyhydride. 21. The method of claim 20,
Wherein the boron oxide is B ₂ O ₃ . The method of claim 7, wherein
And wherein said dielectric material of said second conformal film layer is SiO ₂ . The method of claim 7, wherein
Forming a fourth conformal film layer containing the dielectric material; And
Forming a fifth conformal film layer comprising a dopant source comprising a dopant, on a material adhered to the surface of the substrate,
Introducing the dopant precursor into the reaction chamber;
Adsorbing the dopant precursor on a material adhered to the surface under conditions where the dopant precursor forms an adsorption limiting layer; And
Reacting the adsorbed dopant precursor to form the dopant source, wherein a substantial portion of the fourth conformal film layer is interposed between the third and fifth conformal film layers. Forming the fifth conformal film layer; And
Driving a portion of the dopant from the fifth conformal film layer into the substrate to form a conformal doping profile on the substrate. 24. The method of claim 23,
Forming a sixth film layer that is a capping layer, wherein the capping layer is an outermost film layer of the first to sixth film layers relative to the substrate. . 25. The method of claim 24,
And the cap layer comprises SiO ₂ . 24. The method of claim 23,
Forming the second conformal film layer,
Introducing a reactant into the reaction chamber,
Introducing a dielectric precursor into the reaction chamber,
Adsorbing the dielectric precursor on a material adhered to the substrate under conditions such that the dielectric precursor forms an adsorption limiting layer, and
Reacting the reactant with the dielectric precursor adsorbed to form the dielectric material,
Forming the fourth conformal film layer,
Introducing the reactant into the reaction chamber,
Introducing the dielectric precursor into the reaction chamber;
Adsorbing the dielectric precursor on a material adhered to the substrate under conditions such that the dielectric precursor forms an adsorption limiting layer, and
Reacting the reactant with the dielectric precursor adsorbed to form the dielectric material. The method of claim 26,
Wherein the reactant is an oxidant. The method of claim 27,
Wherein the oxidant is elemental oxygen, nitrous oxide, water, alkyl alcohol, carbon monoxide or carbon dioxide. The method of claim 26,
Wherein said reacting comprises activating with a plasma. The method of claim 26,
The dielectric precursor is
Alkylamino silanes (SiH _x (NR ₂ ) _4-x ) (where x = 1-3 and R comprises an alkyl group) and halosilanes (SiH _x (Y ₂ ) _4-x ) (where x = 1-3, and Y is selected from the group consisting of Cl, Br, and I). The method of claim 26,
Wherein the dielectric precursor is BTBAS. 24. The method of claim 23,
Wherein each of the first to fifth conformal film layers is a monolayer. 24. The method of claim 23,
Wherein each average thickness of the first to fifth conformal film layers is between 0.1 and 2 angstroms. 24. The method of claim 23,
Wherein at least one of the three-dimensional features has a width of about 40 nanometers or less. 24. The method of claim 23,
Wherein the average concentration of dopant on the first to fifth conformal film layers is between about 0.01 and 10 weight percent. A method of doping a patterned substrate in the reaction chamber,
Forming an alloy comprising a dopant source comprising a dopant and a dielectric directly on the surface of the patterned substrate having the three-dimensional features or on a material adhered to the surface, wherein
A dopant precursor and a dielectric precursor are introduced into the reaction chamber and under the conditions under which the dopant precursor and the dielectric precursor form an adsorption-limited layer or on a material adhered to the substrate Cosorbing the dopant precursor and the dielectric precursor, and
Reacting the adsorbed dopant precursor and the adsorbed dielectric precursor to form the dopant source and the alloy having the dopant source; forming the alloy; And
Driving a portion of the dopant from the alloy into the substrate to form a conformal doping profile in the substrate. The method of claim 36,
Forming the alloy further comprises introducing a third reactant into the reaction chamber and reacting the third reactant with the adsorbed dielectric precursor. 39. The method of claim 37,
And the third reactant is an oxidant. The method of claim 38,
Reacting the third reactant with the adsorbed dielectric precursor further comprises activating with a plasma. An apparatus for doping a substrate,
A reaction chamber;
A substrate holder in the reaction chamber;
One or more dopant precursor gas inlets coupled to the reaction chamber;
One or more vacuum pumps;
One or more gas outlets connecting the reaction chamber to the one or more vacuum pumps; And
A controller having machine readable code, the code comprising:
Instructions for determining that a substrate is present on the substrate holder;
Instructions for operating the one or more dopant precursor gas inlets to introduce the dopant precursor into the reaction chamber such that a dopant precursor is adsorbed on the substrate or on a material adhered to the substrate;
Instructions for operating the one or more gas outlets to evacuate nonadsorbed dopant precursor from the reaction chamber, and
Instructions for reacting the adsorbed dopant precursor to form a film layer containing a dopant source. 41. The method of claim 40,
Further comprising an RF generator configured to excite the plasma in the reaction chamber, the instructions for reacting the adsorbed dopant precursor to operate the RF generator to excite the plasma causing the adsorbed dopant precursor to react. And instructions for doping the substrate. 41. The method of claim 40,
And the machine readable code including the controller further includes instructions for driving a portion of the dopant source from the film into the substrate. 43. The method of claim 42,
And a heater configured to heat the film containing the dopant source, instructions for driving a portion of the dopant into the substrate, thermally diffusing a dopant from the film into the substrate by heating the film. and instructions for operating the heater to cause mediated diffusion. A method of doping a patterned substrate in a reaction chamber,
Forming a first conformal film layer comprising a dopant source comprising a arsenic dopant, either directly on the surface of the patterned substrate having three-dimensional features or on a material adhered to the surface,
Introducing a dopant precursor selected from alkylarcin, alkoxyarcin, and aminoarcin chemistries into the reaction chamber;
Adsorbing the dopant precursor directly on the surface or on a material adhered to the surface under conditions in which the dopant precursor forms an adsorption-limited layer; And
Forming the first conformal film layer, comprising reacting the adsorbed dopant precursor to form the dopant source; And
Driving a portion of the dopant from the first conformal film layer into the substrate to form a conformal doping profile on the substrate. 45. The method of claim 44,
Wherein the dopant source is arseno-silicate, arsenic doped silicate glass, arsenic oxide or arsenic oxyhydride. 45. The method of claim 44,
The dopant precursors include arsine, triethylarsenate, trimethyl arsine, triethyl arsine, triphenyl arsine, triphenyl arsine oxide, ethylenebis (diphenyl arsine), tris (dimethylamino) arsine, Or an arsenic containing compound having the formula As (OR) ₃ , wherein R is -CH ₃ or -C ₂ H ₅ . 46. The method of claim 45,
Wherein the dopant source is As ₂ O ₃ and / or As ₂ O ₅ .

Images