US 20060252211 A1
A dielectric film containing a nanolaminate with a hafnium oxide layer and a zirconium oxide layer and a method of fabricating such a dielectric film produce a reliable gate dielectric having an equivalent oxide thickness thinner than attainable using silicon oxide.
1. A method comprising:
forming a layer of hafnium oxide by layering one or more individual monolayers of hafnium oxide;
forming a layer of zirconium oxide on the layer of hafnium oxide to form a nanolaminate including the layer of hafnium oxide and the layer of zirconium oxide, wherein the layer of zirconium oxide is formed by chemical vapor deposition.
2. The method of
3. The method of
4. The method of
5. The method of
6. The method of
7. The method of
8. The method of
9. A method comprising:
forming a dielectric film on a substrate, the dielectric layer having a hafnium oxide/zirconium oxide nanolaminate, the hafnium oxide/zirconium oxide nanolaminate fabricated by
forming a layer of hafnium oxide by layering one or more individual monolayers of hafnium oxide;
forming a layer of zirconium oxide on the layer of hafnium oxide, the layer of zirconium oxide formed by chemical vapor deposition.
10. The method of
11. The method of
12. The method of
13. The method of
14. The method of
15. The method of
16. A method of forming a transistor comprising:
forming first and second source/drain regions in a substrate;
forming a body region between the first and second source/drain regions;
forming a dielectric film above the body region between the first and second source/drain regions, the dielectric film having a nanolaminate containing a layer of hafnium oxide and a layer of zirconium oxide; and
coupling a gate to the dielectric film, wherein forming the nanolaminate includes:
forming the layer of hafnium oxide by layering one or more individual monolayers of hafnium oxide; and
forming the layer of zirconium oxide on the layer of hafnium oxide, wherein the layer of zirconium oxide is formed by chemical vapor deposition.
17. The method of
18. The method of
19. The method of
20. The method of
21. The method of
22. The method of
23. The method of
24. A method of forming a memory comprising:
forming a transistor, the transistor including a dielectric film containing a hafnium oxide/zirconium oxide nanolaminate, the dielectric film formed above a body region between a first source/drain region and a second source/drain region, the hafnium oxide/zirconium oxide nanolaminate formed by:
forming a layer of hafnium oxide by layering one or more individual monolayers of hafnium oxide;
forming a layer of zirconium oxide on the layer of hafnium oxide, the layer of zirconium oxide formed by chemical vapor deposition; and
forming a word line coupled to a gate of the transistor.
25. The method of
26. The method of
27. The method of
28. The method of
29. A method of forming an electronic system comprising:
providing a processor;
coupling a memory to the processor, the memory including a transistor having a dielectric film containing a hafnium oxide/zirconium oxide nanolaminate, the dielectric film formed above a body region between a first source/drain region and a second source/drain region, the hafnium oxide/zirconium oxide nanolaminate formed by:
forming a layer of hafnium oxide by layering one or more individual monolayers of hafnium oxide;
forming a layer of zirconium oxide on the layer of hafnium oxide, the layer of zirconium oxide formed by chemical vapor deposition; and
providing a bus to couple the processor to the memory.
30. The method of
31. The method of
32. The method of
33. The method of
The present application is a Divisional of U.S. application Ser. No. 11/148,505 filed on Jun. 9, 2005 that is a Divisional of U.S. application Ser. No. 10/209,581, filed Jul. 30, 2002, now U.S. Pat. No. 6,921,702, both applications are incorporated herein by reference.
This application is related to the following, co-pending, commonly assigned applications, incorporated herein by reference:
U.S. application Ser. No. 10/163481, entitled: “Atomic Layer-Deposited HfAlO3 Films for Gate Dielectrics;”
U.S. Pat. No. 7,045,430, entitled: “Atomic Layer-Deposited LaAlO3 Films for Gate Dielectrics;”
U.S. application Ser. No. 10/137,058, entitled: “Atomic Layer Deposition and Conversion;”
U.S. application Ser. No. 09/945,535, entitled: “Highly Reliable Amorphous High-K Gate Oxide ZrO2;”
U.S. application Ser. No. 10/137,168, entitled: “Methods, Systems, and Apparatus for Atomic-Layer Deposition of Aluminum Oxides in Integrated Circuits;” and
U.S. Pat. No. 6,852,167, entitled: “Methods, Systems, and Apparatus for Uniform Chemical-Vapor Depositions.”
The invention relates to semiconductor devices and device fabrication. Specifically, the invention relates to gate dielectric layers of transistor devices and their method of fabrication.
The semiconductor device industry has a market driven need to improve speed performance, improve its low static (off-state) power requirements, and adapt to a wide range of power supply and output voltage requirements for it silicon based microelectronic products. In particular, in the fabrication of transistors, there is continuous pressure to reduce the size of devices such as transistors. The ultimate goal is to fabricate increasingly smaller and more reliable integrated circuits (ICs) for use in products such as processor chips, mobile telephones, or memory devices such as DRAMs. The smaller devices are frequently powered by batteries, where there is also pressure to reduce the size of the batteries, and to extend the time between battery charges. This forces the industry to not only design smaller transistors, but to design them to operate reliably with lower power supplies.
Currently, the semiconductor industry relies on the ability to reduce or scale the dimensions of its basic devices, primarily, the silicon based metal-oxide-semiconductor field effect transistor (MOSFET). A common configuration of such a transistor is shown in
In fabricating transistors to be smaller in size and reliably operate on lower power supplies, one important design criteria is the gate dielectric 140. The mainstay for forming the gate dielectric has been silicon dioxide, SiO2. A thermally grown amorphous SiO2 layer provides an electrically and thermodynamically stable material, where the interface of the SiO2 layer with underlying Si provides a high quality interface as well as superior electrical isolation properties. In typical processing, use of SiO2 on Si has provided defect charge densities on the order of 1010/cm2, midgap interface state densities of approximately 1010/cm2 eV, and breakdown voltages in the range of 15 MV/cm. With such qualities, there would be no apparent need to use a material other than SiO2, but increased scaling and other requirements for gate dielectrics create the need to find other dielectric materials to be used for a gate dielectric.
What is needed is an alternate dielectric material for forming a gate dielectric that has a high dielectric constant relative to SiO2, and is thermodynamically stable with respect to silicon such that forming the dielectric on a silicon layer will not result in SiO2 formation, or diffusion of material, such as dopants, into the gate dielectric from the underlying silicon layer.
In the following detailed description of the invention, reference is made to the accompanying drawings which form a part hereof, and in which is shown, by way of illustration, specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. Other embodiments may be utilized and structural, logical, and electrical changes may be made without departing from the scope of the present invention.
The terms wafer and substrate used in the following description include any structure having an exposed surface with which to form the integrated circuit (IC) structure of the invention. The term substrate is understood to include semiconductor wafers. The term substrate is also used to refer to semiconductor structures during processing, and may include other layers that have been fabricated thereupon. Both wafer and substrate include doped and undoped semiconductors, epitaxial semiconductor layers supported by a base semiconductor or insulator, as well as other semiconductor structures well known to one skilled in the art. The term conductor is understood to include semiconductors, and the term insulator or dielectric is defined to include any material that is less electrically conductive than the materials referred to as conductors.
The term “horizontal” as used in this application is defined as a plane parallel to the conventional plane or surface of a wafer or substrate, regardless of the orientation of the wafer or substrate. The term “vertical” refers to a direction perpendicular to the horizontal as defined above. Prepositions, such as “on”, “side” (as in “sidewall”), “higher”, “lower”, “over” and “under” are defined with respect to the conventional plane or surface being on the top surface of the wafer or substrate, regardless of the orientation of the wafer or substrate. The following detailed description is, therefore, not to be taken in a limiting sense, and the scope of the present invention is defined only by the appended claims, along with the full scope of equivalents to which such claims are entitled.
A gate dielectric 140 of
A SiO2 layer of thickness, t, deposited on a Si surface as a gate dielectric will have a teq larger than its thickness, t. This teq results from the capacitance in the surface channel on which the SiO2 is deposited due to the formation of a depletion/inversion region. This depletion/inversion region can result in teq being from 3 to 6 Angstroms (Å) larger than the SiO2 thickness, t. Thus, with the semiconductor industry driving to someday scale the gate dielectric equivalent oxide thickness to under 10 Å, the physical thickness requirement for a SiO2 layer used for a gate dielectric would be need to be approximately 4 to 7 Å.
Additional requirements on a SiO2 layer would depend on the gate electrode used in conjunction with the SiO2 gate dielectric. Using a conventional polysilicon gate would result in an additional increase in teq for the SiO2 layer. This additional thickness could be eliminated by using a metal gate electrode, though metal gates are not currently used in complementary metal-oxide-semiconductor field effect transistor (CMOS) technology. Thus, future devices would be designed towards a physical SiO2 gate dielectric layer of about 5 Å or less. Such a small thickness requirement for a SiO2 oxide layer creates additional problems.
Silicon dioxide is used as a gate dielectric, in part, due to its electrical isolation properties in a SiO2-Si based structure. This electrical isolation is due to the relatively large band gap of SiO2 (8.9 eV) making it a good insulator from electrical conduction. Signification reductions in its band gap would eliminate it as a material for a gate dielectric. As the thickness of a SiO2 layer decreases, the number of atomic layers, or monolayers of the material in the thickness decreases. At a certain thickness, the number of monolayers will be sufficiently small that the SiO2 layer will not have a complete arrangement of atoms as in a larger or bulk layer. As a result of incomplete formation relative to a bulk structure, a thin SiO2 layer of only one or two monolayers will not form a full band gap. The lack of a full band gap in a SiO2 gate dielectric would cause an effective short between an underlying Si channel and an overlying polysilicon gate. This undesirable property sets a limit on the physical thickness to which a SiO2 layer can be scaled. The minimum thickness due to this monolayer effect is thought to be about 7-8 Å. Therefore, for future devices to have a teq less than about 10 Å, other dielectrics than SiO2 need to be considered for use as a gate dielectric.
For a typical dielectric layer used as a gate dielectric, the capacitance is determined as one for a parallel plate capacitance: C=κε0A/t, where κ is the dielectric constant, ε0 is the permittivity of free space, A is the area of the capacitor, and t is the thickness of the dielectric. The thickness, t, of a material is related to teq for a given capacitance with the dielectric constant of SiO2, κox=3.9, associated with teq, as
The thinner equivalent oxide thickness required for lower transistor operating voltages and smaller transistor dimensions may be realized by a significant number of materials, but additional fabricating requirements makes determining a suitable replacement for SiO2 difficult. The current view for the microelectronics industry is still for Si based devices. This requires that the gate dielectric employed be grown on a silicon substrate or silicon layer, which places significant restraints on the substitute dielectric material. During the formation of the dielectric on the silicon layer, there exists the possibility that a small layer of SiO2 could be formed in addition to the desired dielectric. The result would effectively be a dielectric layer consisting of two sublayers in parallel with each other and the silicon layer on which the dielectric is formed. In such a case, the resulting capacitance would be that of two dielectrics in series. As a result, the teq of the dielectric layer would be the sum of the SiO2 thickness and a multiplicative factor of the thickness of the dielectric being formed, written as
In a recent article by G. D. Wilk et al., Journal of Applied Physics, vol. 89: no. 10, pp. 5243-5275 (2001), material properties of high dielectric materials for gate dielectrics were discussed. Among the information disclosed was the viability of Al2O3 as a substitute for SiO2. Al2O3 was disclosed has having favourable properties for use as a gate dielectric such as high band gap, thermodynamic stability on Si up to high temperatures, and an amorphous structure. In addition, Wilk disclosed that forming a layer of Al2O3 on silicon does not result in a SiO2 interfacial layer. However, the dielectric constant of Al2O3 is only 9, where thin layers may have a dielectric constant of about 8 to about 10. Though the dielectric constant of Al2O3 is in an improvement over SiO2, a higher dielectric constant for a gate dielectric is desirable. Other dielectrics and their properties discussed by Wilk include
One of the advantages using SiO2 as a gate dielectric has been that the formation of the SiO2 layer results in an amorphous gate dielectric. Having an amorphous structure for a gate dielectric is advantageous because grain boundaries in polycrystalline gate dielectrics provide high leakage paths. Additionally, grain size and orientation changes throughout a polycrystalline gate dielectric can cause variations in the film's dielectric constant. The abovementioned material properties including crystal structure are for the materials in a bulk form. The materials having the advantage of a high dielectric constants relative to SiO2 also have the disadvantage of a crystalline form, at least in a bulk configuration. The best candidates for replacing SiO2 as a gate dielectric are those with high dielectric constant, which can be fabricated as a thin layer with an amorphous form.
In one embodiment, a method of forming a gate dielectric on a transistor body region includes the formation of HfO2/ZrO2 nanolaminates by atomic layer deposition (ALD) of HfO2 using a HfI4 precursor followed by the formation of ZrO2 on the HfO2 layer. Various embodiments include forming the ZrO2 layer by thermal evaporation followed by krypton/oxygen mixed plasma oxidation, pulsed-laser deposition, or jet-vapor deposition.
A gate dielectric formed as nanolaminates of HfO2/ZrO2 has a larger dielectric constant than silicon dioxide, a relatively small leakage current, and good stability with respect to a silicon based substrate. Embodiments according to the teachings of the present invention include forming transistors, memory devices, and electronic systems having dielectric layers containing nanolaminates of HfO2/ZrO2.
Other embodiments include structures for transistors, memory devices, and electronic systems with gate dielectrics containing nanolaminates of HfO2/ZrO2. Such gate dielectrics provide a significantly thinner equivalent oxide thickness compared with a silicon oxide gate having the same physical thickness. Alternatively, such gate dielectrics provide a significantly thicker physical thickness than a silicon oxide gate dielectric having the same equivalent oxide thickness.
In an embodiment according to the teachings of the present invention, a gate dielectric includes thin layers of HfO2 and ZrO2 forming a nanolaminate. The term “nanolaminate” means a composite film of ultra thin layers of two or more materials in a layered stack, where the layers are alternating layers of materials of the composite film. Typically, nanolaminates have thicknesses of an order of magnitude in the nanometer range. Each individual material layer of the nanolaminate can have thicknesses as low as a monolayer of the material. A nanolaminate of HfO2 and ZrO2 includes at least one thin layer of HfO2, and one thin layer of ZrO2, and is typically written as a nanolaminate of HfO2/ZrO2. In one embodiment, nanolaminates of HfO2/ZrO2 are grown using atomic layer deposition (ALD), also known as atomic layer epitaxy (ALE).
ALD was developed in the early 1970's as a modification of chemical vapor deposition (CVD) and is also called “alternatively pulsed-CVD.” In ALD, gaseous precursors are introduced one at a time to the substrate surface mounted within a reaction chamber (or reactor). This introduction of the gaseous precursors takes the form of pulses of each gaseous precursor. Between the pulses, the reaction chamber is purged with a gas, which in many cases is an inert gas, or evacuated.
In a chemisorption-saturated ALD (CS-ALD) process, during the first pulsing phase, reaction with the substrate occurs with the precursor saturatively chemisorbed at the substrate surface. Subsequent pulsing with a purging gas removes precursor excess from the reaction chamber.
The second pulsing phase introduces another precursor on the substrate where the growth reaction of the desired film takes place. Subsequent to the film growth reaction, reaction byproducts and precursor excess are purged from the reaction chamber. With favourable precursor chemistry where the precursors adsorb and react with each other on the substrate aggressively, one ALD cycle can be preformed in less than one second in properly designed flow type reaction chambers. Typically, precursor pulse times range from about 0.5 sec to about 2 to 3 seconds.
In ALD, the saturation of all the reaction and purging phases makes the growth self-limiting. This self-limiting growth results in large area uniformity and conformality, which has important applications for such cases as planar substrates, deep trenches, and in the processing of porous silicon and high surface area silica and alumina powders. Significantly, ALD provides for controlling film thickness in a straightforward, simple manner by controlling the number of growth cycles.
ALD was originally developed to manufacture luminescent and dielectric films needed in electroluminescent displays. Significant efforts have been made to apply ALD to the growth of doped zinc sulfide and alkaline earth metal sulfide films. Additionally, ALD has been studied for the growth of different epitaxial II-V and II-VI films, nonepitaxial crystalline or amorphous oxide and nitride films and multilayer structures of these. There also has been considerable interest towards the ALD growth of silicon and germanium films, but due to the difficult precursor chemistry, this has not been very successful.
The precursors used in an ALD process may be gaseous, liquid or solid. However, liquid or solid precursors must be volatile. The vapor pressure must be high enough for effective mass transportation. Also, solid and some liquid precursors need to be heated inside the reaction chamber and introduced through heated tubes to the substrates. The necessary vapor pressure must be reached at a temperature below the substrate temperature to avoid the condensation of the precursors on the substrate. Due to the self-limiting growth mechanisms of ALD, relatively low vapor pressure solid precursors can be used though evaporation rates may somewhat vary during the process because of changes in their surface area.
There are several other requirements for precursors used in ALD. The precursors must be thermally stable at the substrate temperature because their decomposition would destroy the surface control and accordingly the advantages of the ALD method which relies on the reactant of the precursor at the substrate surface. Of course, a slight decomposition, if slow compared to the ALD growth, can be tolerated.
The precursors have to chemisorb on or react with the surface, though the interaction between the precursor and the surface as well as the mechanism for the adsorption is different for different precursors. The molecules at the substrate surface must react aggressively with the second precursor to form the desired solid film. Additionally, precursors should not react with the film to cause etching, and precursors should not dissolve in the film. Using highly reactive precursors in ALD contrasts with the selection of precursors for conventional CVD.
The by-products in the reaction must be gaseous in order to allow their easy removal from the reaction chamber. Further, the by-products should not react or adsorb on the surface.
In a reaction sequence ALD (RS-ALD) process, the self-limiting process sequence involves sequential surface chemical reactions. RS-ALD relies on chemistry between a reactive surface and a reactive molecular precursor. In an RS-ALD process, molecular precursors are pulsed into the ALD reaction chamber separately. The metal precursor reaction at the substrate is typically followed by an inert gas pulse to remove excess precursor and by-products from the reaction chamber prior to pulsing the next precursor of the fabrication sequence.
By RS-ALD, films can be layered in equal metered sequences that are all identical in chemical kinetics, deposition per cycle, composition, and thickness. RS-ALD sequences generally deposit less than a full layer per cycle. Typically, a deposition or growth rate of about 0.25 to about 2.00 Å per RS-ALD cycle can be realized.
The advantages of RS-ALD include continuity at an interface, conformality over a substrate, use of low temperature and mildly oxidizing processes, growth thickness dependent solely on the number of cycles performed, and ability to engineer multilayer laminate films with resolution of one to two monolayers. RS-ALD allows for deposition control on the order on monolayers and the ability to deposit monolayers of amorphous films.
RS-ALD processes provide for the formation of nanolaminates. These nanolaminates can be engineered in various forms. In one form, the transition between material layers of the nanolaminate can be made abrupt. In another form, the transition between material layers of the nanolaminate can be constructed with a graded composition. The graded composition can be formed by RS-ALD due its control of the deposition thickness per cycle.
In an embodiment, a layer of HfO2 is formed on a substrate mounted in a reaction chamber using ALD in a repetitive sequence including pulsing a hafnium containing precursor into the reaction chamber followed by pulsing a purging gas, and then pulsing a first oxygen containing precursor into the chamber. In one embodiment using ALD, a layer of HfO2 is formed using HfI4 as a hafnium containing precursor, water vapor as a first oxygen containing precursor, and nitrogen as a purging gas and carrier gas. After forming a HfO2 layer, a ZrO2 layer is formed on the HfO2 layer.
In one embodiment, the layer of ZrO2 is formed by ALD. In particular, a repetitive sequence includes using ZrI4 as a zirconium containing precursor along with a vapor solution of HO2-H2O2 as a second oxygen containing precursor, and nitrogen as a purging gas and carrier gas. In another embodiment, the ZrO2 layer is formed by depositing a layer of zirconium on the HfO2 layer by thermal evaporation, and oxidizing the zirconium layer using a krypton(Kr)/oxygen(O2) mixed plasma to form a HfO2/ZrO2 composite layer. In another embodiment, the ZrO2 layer is formed by pulsed-laser deposition. In yet another embodiment, the ZrO2 layer is formed by jet-vapor deposition.
In one embodiment, precursor gases, in particular HfI4, are used to form the HfO2 layer for the HfO2/ZrO2 nanolaminate films used as a gate dielectric on a transistor body. Alternately, solid or liquid precursors can be used in an appropriately designed reaction chamber. ALD formation of other materials is disclosed in co-pending, commonly assigned U.S. patent application: entitled “Atomic Layer Deposition and Conversion,” attorney docket no. 303.802US1, Ser. No. 10/137,058, and “Methods, Systems, and Apparatus for Atomic-Layer Deposition of Aluminum Oxides in Integrated Circuits,” attorney docket no. 1303.048US1, Ser. No. 10/137,168.
Also included in the ALD system are purging gas sources 261, 262, each of which is coupled to mass-flow controllers 266, 267, respectively. The gas sources 251-254 and the purging gas sources 261-262 are coupled by their associated mass-flow controllers to a common gas line or conduit 270 which is coupled to the gas-distribution fixture 240 inside the reaction chamber 220. Gas conduit 270 is also coupled to vacuum pump, or exhaust pump, 281 by mass-flow controller 286 to remove excess precursor gases, purging gases, and by-product gases at the end of a purging sequence from the gas conduit.
Vacuum pump, or exhaust pump, 282 is coupled by mass-flow controller 287 to remove excess precursor gases, purging gases, and by-product gases at the end of a purging sequence from the reaction chamber 220. For convenience, control displays, mounting apparatus, temperature sensing devices, substrate maneuvering apparatus, and necessary electrical connections as are known to those skilled in the art are not shown in
The use, construction and fundamental operation of reaction chambers for deposition of films are understood by those of ordinary skill in the art of semiconductor fabrication. The present invention man be practiced on a variety of such reaction chambers without undue experimentation. Furthermore, one of ordinary skill in the art will comprehend the necessary detection, measurement, and control techniques in the art of semiconductor fabrication upon reading the disclosure.
Forming the HfO2 layer on a substrate by atomic layer deposition involves using a deposition sequence including pulsing the HfI4 precursor into the reaction chamber, followed by pulsing a purging gas, pulsing a first oxygen containing precursor, and pulsing the purging gas. In one embodiment, the first oxygen precursor is water vapor. Each precursor is pulsed for a short time ranging from 0.5 seconds to two or three seconds. A purging gas such as nitrogen is pulsed for a longer period such as five to fifteen seconds to insure that all excess precursor gases and by-products are removed from the reaction chamber. Pulsing times are selected to enable the controlled growth of the HfO2 layer on a one to two monolayer basis. For a fixed ALD sequence or cycle, including fixed pulsing times and substrate temperatures, the HfO2 layer growth rate is at a relatively fixed rate, where a desired thickness of the HfO2 layer is obtained by performing the ALD sequence for a predetermined number of cycles.
At block 405, a substrate is prepared. The substrate used for forming a transistor is typically a silicon or silicon containing material. In other embodiments, germanium, gallium arsenide, silicon-on-sapphire substrates, or other suitable substrates may be used. This preparation process includes cleaning of the substrate 210 and forming layers and regions of the substrate, such as drains and sources of a metal oxide semiconductor (MOS) transistor, prior to forming a gate dielectric. The sequencing of the formation of the regions of the transistor being processed follows typical sequencing that is generally performed in the fabrication of a MOS transistor as is well known to those skilled in the art. Included in the processing prior to forming a gate dielectric is the masking of substrate regions to be protected during the gate dielectric formation, as is typically performed in MOS fabrication. In this embodiment, the unmasked region includes a body region of a transistor, however one skilled in the art will recognize that other semiconductor device structures may utilize this process. Additionally, the substrate 210 in its ready for processing form is conveyed into a position in reaction chamber 220 for ALD processing.
At block 410, a precursor containing hafnium is pulsed into reaction chamber 220. In particular, HfI4 is used as a source material. The HfI4 is pulsed into reaction chamber 220 through the gas-distribution fixture 240 onto substrate 210. The flow of the HfI4 is controlled by mass-flow controller 256 from gas source 251. In one embodiment, the substrate temperature is maintained between about 225° C. and about 500° C. In another embodiment, the substrate temperature is maintained between about 250° C. and about 325° C. The lower temperature allows for forming a dielectric film suited for use as a gate dielectric, since an amorphous layer tends to more readily form at lower processing temperatures. The HfI4 reacts with the surface of the substrate 210 in the desired region defined by the unmasked areas of the substrate 210.
At block 415, a first purging gas is pulsed into the reaction chamber 220. In particular, pure nitrogen with a purity greater than 99.99% is used as a purging gas for HfI4. The nitrogen flow is controlled by mass-flow controller 266 from the purging gas source 261 into the gas conduit 270. Using the pure nitrogen purge avoids overlap of the precursor pulses and possible gas phase reactions. A nitrogen gas can also be used as a carrier gas for the precursors. Following the purge, a first oxygen containing precursor is pulsed into the reaction chamber 220, at block 420. For the hafnium sequence using HfI4 as the precursor, water vapor is selected as the precursor acting as an oxidizing reactant to form a HfO2 on the substrate 210. Alternately, a vapor solution of H2O-H2O2 can be used as the oxygen containing precursor. The water vapor is pulsed into the reaction chamber 220 through gas conduit 270 from gas source 252 by mass-flow controller 257. The water vapor aggressively reacts at the surface of substrate 210.
Following the pulsing of oxidizing reactant water vapor, the first purging gas is injected into the reaction chamber 220, at block 425. In the HfI4/water vapor sequence, pure nitrogen gas is used to purge the reaction chamber after pulsing each precursor gas. Excess precursor gas, and reaction by-products are removed from the system by the purge gas in conjunction with the exhausting of the reaction chamber 220 using vacuum pump 282 through mass-flow controller 287, and exhausting of the gas conduit 270 by the vacuum pump 281 through mass-flow controller 286.
During the HfI4/water vapor sequence, the substrate is held between about 250° C. and about 325° C. by the heating element 230. In other embodiments the substrate is held between about 225° C. and 500° C. The HfI4 pulse time ranges from about 1.0 sec to about 2.0 sec. After the HfI4 pulse, the hafnium sequence continues with a purge pulse followed by a water vapor pulse followed by a purge pulse. In one embodiment, performing a purge pulse followed by a water vapor pulse followed by a purge pulse takes about 2 seconds. In another embodiment, each pulse in the hafnium sequence has a 2 second pulse period. In another embodiment, the pulse periods for the precursors are 2 seconds, while the purge gas pulse period ranges from five second to twenty seconds.
At block 430, a determination is made as to whether a desired thickness of the HfO2 layer has been formed. The thickness of a HfO2 film after one cycle is determined by a fixed growth rate for the pulsing periods and precursors used in the hafnium sequence, set at a value such as N nm/cycle. For a desired HfO2 film thickness, t, in an application such as forming a gate dielectric of a MOS transistor, the ALD process should be repeated for t/N cycles. The desired thickness should be attained after t/N cycles. If less than t/N cycles have been completed, the process starts over at block 410 with the pulsing of the precursor containing hafnium, which in the embodiment discussed above is a HfI4 gas. If t/N cycles have completed, no further ALD processing of HfO2 is required and the HfO2 layer is ready to be formed as a composite with a ZrO2 layer.
At block 435, a precursor containing zirconium is pulsed into the reaction chamber 220. In one embodiment, ZrI4 is used as the zirconium containing precursor. In another embodiment, ZrCl4 is used as the zirconium containing precursor. The ZrI4 is evaporated from a containment area held at about 250° C. in gas source 253. It is pulsed to the surface of the substrate 210 through gas-distribution fixture 240 from gas source 253 by mass-flow controller 258. The ZrI4 is introduced onto the HfO2 layer that was formed during the HfI4/water vapor sequence.
At block 440, a second purging gas is introduced into the system. For a ZrI4 precursor, nitrogen gas is used as a purging and carrier gas. The nitrogen flow is controlled by mass-flow controller 267 from the purging gas source 262 into the gas conduit 270 and subsequently into the reaction chamber 220. Following the nitrogen purge, at block 445, a second oxygen containing precursor is pulsed into the reaction chamber 220. For the zirconium sequence using ZrI4 as the precursor, a vapor solution of H2O-H2O2 is selected as the precursor acting as an oxidizing reactant to interact with the zirconium deposited on the HfO2 layer on the substrate 210. The H2O-H2O2 vapor solution is pulsed into the reaction chamber 220 through gas conduit 270 from gas source 254, held at about room temperature, by mass-flow controller 259. The H2O-H2O2 vapor solution aggressively reacts at the surface of substrate 210 to form a ZrO2 layer.
Following the pulsing of the H2O-H2O2 vapor solution acting as an oxidizing reactant, the nitrogen purging gas is injected into the reaction chamber 200, at block 450. In the ZrI4/H2-H2O2 vapor solution sequence, nitrogen gas is used to purge the reaction chamber after pulsing each precursor gas. In another embodiment, argon gas is used as the purging gas. Excess precursor gas, and reaction by-products are removed from the system by the purge gas in conjunction with the exhausting of the reaction chamber 220 using vacuum pump 282 through mass-flow controller 287, and exhausting of the gas conduit 270 by the vacuum pump 281 through mass-flow controller 286.
During the ZrI4/H2O-H2O2 vapor solution sequence, the substrate is held between about 250° C. and about 325° C. by the heating element 230. In other embodiments, the substrate is held between about 275° C. and about 500° C. In one embodiment, the process pressure is maintained at about 250 Pa during the zirconium sequence. Pulse times for the ZrI4 and the H2O-H2O2 vapor solution were about 2 sec for both precursors, with purging pulse times of about 2 secs.
At 455, similar to the HfO2 layer formation, a determination is made as to whether the ZrO2 layer has the desired thickness by determining if a desired number of zirconium cycles have been performed. If the number of zirconium cycles performed is less than the number needed to form the desired thickness, the zirconium containing precursor is pulsed into the reaction chamber, at block 435, and the process continues. If the desired number of zirconium cycles has been performed, this completes not only the ZrI4/H2O-H2O2 vapor solution sequence, but it also completes a hafnium sequence/zirconium sequence cycle forming a HfO2/ZrO2 nanolaminate.
Upon completing the formation of the HfO2/ZrO2 nanolaminate, the nanolaminate can be annealed. The annealing can be performed at a temperature between about 300° C. and about 800° C. in an inert or nitrogen atmosphere.
At block 460, after forming the HfO2/ZrO2 nanolaminate, processing the device containing the HfO2/ZrO2 nanolaminate is completed. In one embodiment, completing the device includes completing the formation of a transistor. Alternately, completing the process includes completing the construction of a memory device having a array with access transistors formed with gate dielectrics containing HfO2/ZrO2 nanolaminates. Further, in another embodiment, completing the process includes the formation of an electronic system including an information handling device that uses electronic devices with transistors formed with gate dielectrics containing HfO2/ZrO2 nanolaminates. Typically, information handling devices such as computers include many memory devices, having many access transistors.
In one embodiment, a HfO2/ZrO2 nanolaminate includes one HfO2 layer and one HfO2/ZrO2 layer. The completed HfO2/ZrO2 nanolaminate has a thickness in which the thickness of the HfO2 layer is about one-half the thickness of the completed HfO2/ZrO2 nanolaminate. In another embodiment, a completed HfO2/ZrO2 nanolaminate includes multiple alternating layers of HfO2 and ZrO2, which requires that at block 455, once a given ZrO2 layer has been formed with a desired thickness, a hafnium sequence is then started at block 410. This process, proceeding from completing the zirconium sequence at block 455 to starting the hafnium sequence at block 410, continues until the desired number of alternating layers of HfO2 and ZrO2 have been formed. The HfO2/ZrO2 nanolaminate formation begins with forming a HfO2 layer, but may end with forming ZrO2 layer or a HfO2 layer. ALD provides for the engineering of a HfO2/ZrO2 nanolaminate. For example, nanolaminates can be formed with n number of HfO2/ZrO2 composite layers where the HfO2 layer is formed with x number of hafnium cycles and y number of zirconium cycles. Alternately, nanolaminates can be formed with n number of HfO2/ZrO2 composite layers where the first composite layer has a HfO2 layer formed with x1 number of hafnium cycles and y1 number of zirconium cycles, a second composite layer has a HfO2 layer formed with x2 number of hafnium cycles and y2 number of zirconium cycles, extended to the nth composite layer having a HfO2 layer formed with xn number of hafnium cycles and yn number of zirconium cycles. Such tailoring of the HfO2/ZrO2 nanolaminate provides for forming dielectric films with a designed physical thickness, t, and equivalent oxide thickness, teq.
In the hafnium sequence and in the zirconium sequence, pulsing each precursor into the reaction chamber is controlled for a predetermined period, the predetermined period being individually controlled for each precursor pulsed into the reaction chamber. Additionally, the substrate is maintained at a selected temperature for forming each layer, where the selected temperature set independently for forming each layer.
In a recent article by O. Sneh et al., Thin Solid Films, vol. 402, pp. 248-261 (2002), atomic layer deposition of thin films was discussed. The article noted that the growth rate for HfO2 is, typically, about 0.8 Å/cycle. Similarly, in a recent article by K. Kukli et al., Journal of the Electrochemical Society, vol. 148, no. 12, pp. F227-F232 (2001), dealing with ZrO2 formed by ALD using ZrI4, it was noted that at about a growth temperature of about 300° C., ZrO2 growth rate was about 0.075 nm/cycle. Thus, in the embodiments for forming HfO2/ZrO2 nanolaminates using ALD for all composite layers, each material layer can be grown at about 0.75-0.80 Å/cycle.
During the evaporation process, the electron gun 663 generates an electron beam 664 that hits target 661. In one embodiment, target 661 contains a zirconium metal source, which is evaporated due to the impact of the electron beam 664. The evaporated material 668 is then distributed throughout the chamber 660. A layer of zirconium is grown forming a film 640 on the surface of the HfO2 layer 633 on substrate 610, which is maintained at a temperature between 150° C. and 200° C. The growth rate can vary with a typical rate of 0.1 Å/s. After depositing a zirconium layer on the HfO2 layer 633, the zirconium layer is oxidized.
The evaporation chamber 660 can be included as part of an overall processing system including ALD system 200 of
In an embodiment, the processing variables for the mixed plasma oxidation include a low ion bombardment energy of less than 7 eV, a high plasma density above 1012/cm3 and a low electron temperature below 1.3 eV. In another embodiment, the substrate temperature is approximately 400° C. In another embodiment, a mixed gas of 3% oxygen with the balance being krypton at a pressure of 1 Torr is used. In one embodiment, a microwave power density of 5 W/cm2 is used. The oxidation process provides a growth rate of 1.5 nm/min.
The low temperature mixed plasma oxidation process described above allows the deposited layer to be oxidized at a low temperature. The mixed plasma process in one embodiment is performed at approximately 400° C. in contrast to prior thermal oxidation processes that are performed at approximately 1000° C.
In one embodiment, a substrate temperature is maintained between about 200° C. to about 800° C. during pulsed-laser deposition. A beam from a laser source such as a excimer laser is focused on a rotating zirconium target source in a deposition chamber with an O2 pressure of about 0.2 Torr to form a ZrO2 layer on a HfO2 layer. Other laser sources and configurations can be used as is known by those skilled in the art.
In one embodiment, the jet-vapor deposition zirconium and oxygen vapors are directed to the HfO2 layer out of source nozzles by supersonic Ar jets. Using jet-vapor deposition in a low pressure atmosphere allows for forming the ZrO2 layer at at room temperature. In one embodiment, annealing is performed subsequent to forming the ZrO2 layer at a about 800° C. In one embodiment, the annealing is performed using a nitrogen rapid thermal annealing (RTA). The annealing can be performed after each ZrO2 layer is formed in the composite of alternating layers of the HfO2/ZrO2 nanolaminate and/or at the completion of the HfO2/ZrO2 nanolaminate.
In each of the various embodiments for forming ZrO2 layers, the HfO2/ZrO2 nanolaminates can be annealed in a temperature range from between 300° C. to 800° C. Typically the annealing is for a short time and in performed in a nitrogen atmosphere or in some other inert atmosphere.
Bulk layers of HfO2 and bulk layers of ZrO2 both have a dielectric constant of about 25. Consequently, a material film composed of bulk layers of HfO2 and ZrO2 will also have a dielectric constant of about 25. However, thin layers of a material, typically, have dielectric constants somewhat less than their bulk counterparts. The reduced value of the dielectric constants for ultra thin material films is due in part to the formation of an interfacial layer between the material film and the substrate. Some materials formed on silicon substrates form a SiO2 interfacial layer, while other materials form an silicide interfacial layer. The material silicide in many cases will have a dielectric greater than SiO2, but less than the bulk material dielectric constant. ZrO2 formed on silicon substrates may result in an interfacial region where silicon diffuses through a layer of ZrO2 to form a poly-silicon/ ZrO2/silicon interfacial region, as reported by C. H. Lee et al., IEDM 2000, 27-30 (2000). Further, nanolaminates of ZrO2/HfO2 were reported to have SiO2 interfacial layer when formed by ALD using ZrCl4 and HfCl4 precursors. See H. Zhang et al., Journal of the Electrochemical Society, vol. 148, no. 4, pp. F63-F66 (2001). To eliminate the SiO2 interfacial layer, Zhang et al. grew ZrO2/HfO2 nanolaminates on nitrated Si substrates producing dielectric constants ranging from 9 to 14 with low leakage currents ranging from 2.2×10−6 to 1.2×10−8 A/cm3 at 1 MV/cm.
In the various embodiments according to the teachings of the present invention, HfO2/ZrO2 nanolaminates are formed by ALD of HfO2 on substrates using a HfI4 precursor. Subsequently, a layer of ZrO2 is formed on the HfO2 layer by various deposition techniques. These HfO2/ZrO2 nanolaminates form a stable interface with a silicon substrate. Using ALD, the size and effect of interfacial layer between the silicon substrate and the first HfO2 layer will depend on the reactivity of the HfO2 in forming an abrupt transition from silicon surface to HfO2 layer. Consequently, dielectric films containing HfO2/ZrO2 nanolaminates can have dielectric constants ranging from 9 or 10 to 25. Additionally, forming the HfO2 layer at relatively low temperatures provides a means for enabling the formation of HfO2/ZrO2 nanolaminates that are amorphous.
Another factor setting a lower limit for the scaling of a dielectric layer is the number of monolayers of the dielectric structure necessary to develop a full band gap such that good insulation is maintained between an underlying silicon layer and an overlying conductive layer on the dielectric layer or film. This requirement is necessary to avoid possible short circuit effects between the underlying silicon layer and the overlying conductive layer used. In one embodiment, for several HfO2 monolayers and several ZrO2 monolayers forming a nanolaminate, an expected lower limit for the physical thickness of a dielectric layer grown by forming HfO2/ZrO2 nanolaminates is anticipated to be in about the 2-4 nm range. Consequently, typical dielectric layers or films can be grown by forming HfO2/ZrO2 nanolaminates having physical thickness in the range of 4 to 10 nm. HfO2 used as the initial layer is expected to provide excellent overall results with respect to reliability, current leakage, and ultra-thin teq. Further, using ALD for processing all layers of a HfO2/ZrO2 nanolaminate, the transitions between such oxide layers can be engineered to be abrupt or graded. Thus, the number of layers used, the thickness of each layer, and the nature of the interface between each layer can be engineered to provide the desired electrical characteristics.
With HfO2 layers formed by ALD and ZrO2 layers formed according to one of the various embodiments described herein, HfO2/ZrO2 nanolaminates can have a wide range of thicknesses and dielectric constants. The physical thicknesses can range from about 2 nm to about 10 nm with typical thickness ranging from about 4 nm to about 10 nm. Such layers have an effective dielectric constant ranging from 9 or 10 to 25. The expected teq ranges for various effective dielectric constants are shown in the following:
As mentioned, the lower limit on the scaling of a layer containing HfO2/ZrO2 nanolaminates depends on the monolayers of the film necessary to develop a full band gap such that good insulation is maintained between an underlying silicon layer and an overlying conductive layer to the HfO2/ZrO2 nanolaminate film. From above, it is apparent that a film containing HfO2/ZrO2 nanolaminates can be attained with a teq ranging from 3 Å to 12 Å. Further, a dielectric film with completely formed band structures and monolayer formations can provide a teq significantly less than 2 or 3 Å.
The novel process described above provides significant advantages by performing atomic layer deposition of HfO2/ZrO2 in a hafnium sequence using HfI4 precursors followed by the formation of a ZrO2 layer on the HfO2 layer. Further, by independently controlling the various parameters for each sequence a gate dielectric with a selected dielectric constant can be formed. Additionally, the novel process can be implemented to form transistors, memory devices, and information handling devices. With careful preparation and engineering of the HfO2/ZrO2 nanolaminates limiting the size of interfacial regions, a teq down to 2.5 Å or lower is anticipated.
A transistor 100 as depicted in
A gate is formed over the gate dielectric 140. Typically, forming the gate includes forming a polysilicon layer, though a metal gate can be formed in an alternative process. Forming the substrate, source/region regions, and the gate is performed using standard processes known to those skilled in the art. Additionally, the sequencing of the various elements of the process for forming a transistor is conducted with standard fabrication processes, also as known to those skilled in the art.
Embodiments of the method of forming HfO2/ZrO2 nanolaminates as a gate dielectric can be applied to other transistor structures having dielectric layers. For example,
Transistors created by the methods described above may be implemented into memory devices and electronic systems including information handling devices. Information handling devices having a dielectric layer containing HfO2/ZrO2 nanolaminates can be constructed using various embodiments of the methods described above. Such information devices can include wireless systems, telecommunication systems, and computers. An embodiment of a computer having a dielectric layer containing HfO2/ZrO2 nanolaminates is shown in
Microprocessor 1306 produces control and address signals to control the exchange of data between memory bus circuit 1308 and microprocessor 1306 and between memory bus circuit 1308 and peripheral circuitry 1310. This exchange of data is accomplished over high speed memory bus 1320 and over high speed I/O bus 1322.
Coupled to memory bus 1320 are a plurality of memory slots 1312(a-n) which receive memory devices well known to those skilled in the art. For example, single in-line memory modules (SIMMs) and dual in-line memory modules (DIMMs) may be used in the implementation of the present invention.
These memory devices can be produced in a variety of designs which provide different methods of reading from and writing to the dynamic memory cells of memory slots 1312. One such method is the page mode operation. Page mode operations in a DRAM are defined by the method of accessing a row of a memory cell arrays and randomly accessing different columns of the array. Data stored at the row and column intersection can be read and output while that column is accessed. Page mode DRAMs require access steps which limit the communication speed of memory circuit 1308.
An alternate type of device is the extended data output (EDO) memory which allows data stored at a memory array address to be available as output after the addressed column has been closed. This memory can increase some communication speeds by allowing shorter access signals without reducing the time in which memory output data is available on memory bus 1320. Other alternative types of devices include SDRAM, DDR SDRAM, SLDRAM and Direct RDRAM as well as others such as SRAM or Flash memories.
Control, address and data information provided over memory bus 1320 is further represented by individual inputs to DRAM 1400, as shown in
As is well known in the art, DRAM 1400 includes memory array 1410 which in turn comprises rows and columns of addressable memory cells. Each memory cell in a row is coupled to a common word line. The word line is coupled to gates of individual transistors, where at least one transistor has a gate coupled to a gate dielectric containing HfO2/ZrO2 nanolaminates in accordance with the method and structure previously described above. Additionally, each memory cell in a column is coupled to a common bit line. Each cell in memory array 1410 includes a storage capacitor and an access transistor as is conventional in the art.
DRAM 1400 interfaces with, for example, microprocessor 1306 through address lines 1404 and data lines 1402. Alternatively, DRAM 1400 may interface with a DRAM controller, a micro-controller, a chip set or other electronic system. Microprocessor 1306 also provides a number of control signals to DRAM 1400, including but not limited to, row and column address strobe signals RAS and CAS, write enable signal WE, an output enable signal OE and other conventional control signals.
Row address buffer 1412 and row decoder 1414 receive and decode row addresses from row address signals provided on address lines 1404 by microprocessor 1306. Each unique row address corresponds to a row of cells in memory array 1410. Row decoder 1414 includes a word line driver, an address decoder tree, and circuitry which translates a given row address received from row address buffers 1412 and selectively activates the appropriate word line of memory array 1410 via the word line drivers.
Column address buffer 1416 and column decoder 1418 receive and decode column address signals provided on address lines 1404. Column decoder 1418 also determines when a column is defective and the address of a replacement column. Column decoder 1418 is coupled to sense amplifiers 1420. Sense amplifiers 1420 are coupled to complementary pairs of bit lines of memory array 1410.
Sense amplifiers 1420 are coupled to data-in buffers 1422 and data-out buffers 1424. Data-in buffers 1422 and data-out buffers 1424 are coupled to data lines 1402. During a write operation, data lines 1402 provide data to data-in buffers 1422. Sense amplifier 1420 receives data from data-in buffers 1422 and stores the data in memory array 1410 as a charge on a capacitor of a cell at an address specified on address lines 1404.
During a read operation, DRAM 1400 transfers data to microprocessor 1306 from memory array 1410. Complementary bit lines for the accessed cell are equilibrated during a precharge operation to a reference voltage provided by an equilibration circuit and a reference voltage supply. The charge stored in the accessed cell is then shared with the associated bit lines. A sense amplifier of sense amplifiers 1420 detects and amplifies a difference in voltage between the complementary bit lines. The sense amplifier passes the amplified voltage to data-out buffers 1424.
Control logic 1406 is used to control the many available functions of DRAM 1400. In addition, various control circuits and signals not detailed herein initiate and synchronize DRAM 1400 operation as known to those skilled in the art. As stated above, the description of DRAM 1400 has been simplified for purposes of illustrating the present invention and is not intended to be a complete description of all the features of a DRAM. Those skilled in the art will recognize that a wide variety of memory devices, including but not limited to, SDRAMs, SLDRAMs, RDRAMs and other DRAMs and SRAMs, VRAMs and EEPROMs, may be used in the implementation of the present invention. The DRAM implementation described herein is illustrative only and not intended to be exclusive or limiting.
A gate dielectric containing HfO2/ZrO2 nanolaminates and a method of fabricating such a gate produces a reliable gate dielectric having an equivalent oxide thickness thinner than attainable using SiO2. Gate dielectrics containing HfO2/ZrO2 nanolaminates formed using the methods described herein are thermodynamically stable such that the gate dielectrics formed will have minimal reactions with a silicon substrate or other structures during processing.
Transistors, higher level ICs or devices, and systems are constructed utilizing the novel process for forming a gate dielectric having an ultra thin equivalent oxide thickness, teq. Gate dielectric layers or films containing HfO2/ZrO2 nanolaminates are formed having a high dielectric constant (κ), where the gate dielectrics are capable of a teq thinner than 10 Å, thinner than the expected limit for SiO2 gate dielectrics. At the same time, the physical thickness of the HfO2/ZrO2 nanolaminates is much larger than the SiO2 thickness associated with the teq limit of SiO2. Forming the larger thickness provides advantages in processing the gate dielectric. Further, HfO2/ZrO2 nanolaminates processed in relatively low temperatures can provide amorphous dielectric films having relatively low leakage current for use as dielectric layers in electronic devices and systems.
Although specific embodiments have been illustrated and described herein, it will be appreciated by those of ordinary skill in the art that any arrangement which is calculated to achieve the same purpose may be substituted for the specific embodiment shown. This application is intended to cover any adaptations or variations of the present invention. It is to be understood that the above description is intended to be illustrative, and not restrictive. Combinations of the above embodiments, and other embodiments will be apparent to those of skill in the art upon reviewing the above description. The scope of the invention includes any other applications in which the above structures and fabrication methods are used. The scope of the invention should be determined with reference to the appended claims, along with the full scope of equivalents to which such claims are entitled.