Many modern day electronic devices include non-volatile memory. Non-volatile memory is electronic memory that is able to store data when powered and also in the absence of power. A promising candidate for the next generation of non-volatile memory is ferroelectric random-access memory (FeRAM). FeRAM has a relatively simple structure and is compatible with complementary metal-oxide-semiconductor (CMOS) logic fabrication processes.
Aspects of the present disclosure are best understood from the following detailed description when read with the accompanying figures. It is noted that, in accordance with the standard practice in the industry, various features are not drawn to scale. In fact, the dimensions of the various features may be arbitrarily increased or reduced for clarity of discussion.
The following disclosure provides many different embodiments, or examples, for implementing different features of the provided subject matter. Specific examples of components and arrangements are described below to simplify the present disclosure. These are, of course, merely examples and are not intended to be limiting. For example, the formation of a first feature over or on a second feature in the description that follows may include embodiments in which the first and second features are formed in direct contact, and may also include embodiments in which additional features may be formed between the first and second features, such that the first and second features may not be in direct contact. In addition, the present disclosure may repeat reference numerals and/or letters in the various examples. This repetition is for the purpose of simplicity and clarity and does not in itself dictate a relationship between the various embodiments and/or configurations discussed.
Further, spatially relative terms, such as “beneath,” “below,” “lower,” “above,” “upper” and the like, may be used herein for ease of description to describe one element or feature's relationship to another element(s) or feature(s) as illustrated in the figures. The spatially relative terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. The apparatus may be otherwise oriented (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein may likewise be interpreted accordingly.
A ferroelectric field-effect transistor (FeFET) device is a type of ferroelectric device comprising ferroelectric material arranged between a conductive gate structure and a channel region, where the channel region is disposed between a source region and a drain region. During operation of the FeFET device, applying a gate voltage to the gate structure generates an electric field that establishes a dipole moment within the ferroelectric material. Depending on a value of the gate voltage, a direction of the dipole moment (e.g., a polarization) may be in one various discrete directions. Since a threshold voltage (e.g., a minimum gate-to-source voltage that forms a conductive path between the source region and the drain region) of a FeFET device is dependent upon the polarization within the ferroelectric material, the different polarizations effectively split the threshold voltage of the FeFET device into distinct values corresponding to different data states. The polarization states are maintained when the gate voltage/electric field is removed.
A ferroelectric device that includes a ferroelectric dominant lattice structure has only one positive polarization state and only one negative state, for a total of two states which may be used to store one bit in a single memory cell. In contrast, a ferroelectric device that includes an anti-ferroelectric dominant lattice structure has two positive polarization states and two negative polarization states, for a total of four states which may be used to store two digital bits in a single memory cell. Thus, adopting anti-ferroelectric dominant material can allow for twice the amount of data to be stored in the same footprint, thereby reducing memory sizes and/or increasing data storage densities compared to memory cells made of solely ferroelectric dominant material.
One way to fabricate ferroelectric devices with an anti-ferroelectric dominant lattice structure is to increase a percentage of dopants in a ferroelectric material. For example, for a hafnium (Hf) zirconium (Zr) oxide (O) (HZO) (e.g., Hf1-xZrxO) ferroelectric device, increasing a Zr molar/dopant concentration (e.g., x) to be greater than 70% can achieve a tetragonal lattice structure with anti-ferroelectric properties. This makes the HZO recipe more complicated and may lead to uneven Hf/Zr distribution in the HZO film. In addition, Zr-rich HZO shows a strong wake-up effect where the polarization increases over time as the number of write/read electric field cycles is increased. The wake-up effect negatively affects endurance and causes performance instability. Further, ferroelectric devices with high Zr dopant concentrations can suffer from undesirable current leakage, which ultimately degrades the performance and/or stability of the FeFET devices. This current leakage can arise, for example, from device fatigue as the devices ages, and/or can arise as defects are generated over time during polarization switching. For instance, if there are a large number of oxygen defects in the oxide semiconductor and/or ferroelectric film, this can lead to a potential pathway for device breakdown.
Accordingly, in view of the above, various embodiments of the present disclosure relate to an integrated chip having a FeFET device comprising a ferroelectric structure that includes a ferroelectric layer and an anti-ferroelectric layer with improved endurance properties. In particular, the ferroelectric layer exhibits a dominant orthorhombic crystalline structure that exhibits ferroelectric properties, while the anti-ferroelectric layer exhibits a dominant tetragonal crystalline structure that exhibits anti-ferroelectric properties. An interlayer separates the ferroelectric layer from the anti-ferroelectric layer, and enhances tetragonal crystalline grain growth in the anti-ferroelectric layer. The interlayer also acts to suppress current leakage, for example by promoting the formation of large grains in the ferroelectric structure to help suppress current leakage through defects.
In some embodiments, the oxide semiconductor 106 comprises one or more of indium gallium oxide (IGO), indium gallium zinc oxide (IGZO), indium gallium zinc tin oxide, indium tungsten oxide (IWO), indium tungsten zinc oxide, indium zinc oxide (IZ), zinc oxide, and/or Sn-doped oxide semiconductor materials. In some cases, the oxide semiconductor 106 comprises InxGayZnzMO, where M can be Ti, Al, Ag, W, Ce, or Sn; and x, y, and Z each range between 0 and 1. The gate structure 102, source region 108, and drain region 109 can each have a thickness of between 50 angstroms and 500 angstroms, and can comprise a metal, such as tantalum nitride (TaN), titanium (Ti), Tungsten (W), Titanium nitride (TiN), molybdenum, copper, gold, zinc, aluminum, or the like. In various embodiments, the gate structure 102 corresponds to a wordline of a FeRAM device (see e.g.,
Compared to related FeFET devices which include only a single ferroelectric layer, the ferroelectric structure 104 of
The ferroelectric layer 110 can comprise hafnium oxide, such as Hf1-xZrxO, or similar, and the anti-ferroelectric layer 112 can also comprise hafnium oxide, such as Hf1-yZryO, where x and y are molar concentrations of Zr and are different from one another. In some cases, x is less than 0.70, which can provide homogenous Zr distribution through the HZO lattice and reduce wake-up effect. For example, x can range between 0 and 0.5, and y can range between 0.5 and 1. The ferroelectric layer 110 can be thicker than the anti-ferroelectric layer 112, can have the same thickness as the anti-ferroelectric layer 112, or can be thinner than the anti-ferroelectric layer 112.
The ferroelectric layer 110 can have a lattice with a different orientation than that of the anti-ferroelectric layer 112. For example, the ferroelectric layer 110 can have a predominantly orthorhombic crystal lattice while the anti-ferroelectric layer 112 has a predominantly tetragonal crystal lattice. Thus, the ferroelectric layer has a first percentage of tetragonal crystalline lattice per unit volume and the anti-ferroelectric layer has a second percentage of tetragonal crystalline lattice per unit volume, wherein the second percentage is greater than the first percentage. Further, the ferroelectric layer 110 can be more than 50% of orthorhombic crystalline lattice per unit volume, more than 70% of orthorhombic crystalline lattice per unit volume, or even more than 90% of orthorhombic crystalline lattice per unit volume; and the anti-ferroelectric layer 112 can have more than 50% of tetragonal crystalline lattice per unit volume, more than 70% of tetragonal crystalline lattice per unit volume, or even more than 90% of tetragonal crystalline lattice per unit volume. In some cases, a percentage of orthorhombic crystalline lattice per unit volume in the ferroelectric layer 110 is greater than a percentage of tetragonal crystalline lattice per unit volume in the anti-ferroelectric layer 112. For instance, in some examples, a ratio of the percentage of orthorhombic crystal per unit volume in the ferroelectric layer 110 to the percentage of tetragonal crystal per unit volume in the anti-ferroelectric layer 112 is in a range of 15:1 to 25:1. Such ratios of crystalline lattice structures can help provide the ferroelectric and anti-ferroelectric properties.
In some cases, the interlayer 114 comprises a metal oxide, such as titanium oxide, aluminum oxide, magnesium oxide, hafnium oxide, or indium oxide; or a metal, such as titanium, platinum, gold, or nickel. In some cases, the interlayer has a coefficient of thermal expansion of 1×10-6 meters (m)/Kelvin (K) to 1×10-4/K. The interlayer 114 can have a thickness of less than 1 nm, which promotes high quality crystal growth on surfaces of the interlayer 114.
The interlayer 114 provides a break between crystalline lattices of the ferroelectric layer 110 and anti-ferroelectric layer 112, and promotes a dominant phase of the ferroelectric structure 104 from the orthorhombic phase to the tetragonal phase. By properly introducing the interlayer 114, the anti-ferroelectric properties are induced in the ferroelectric structure 104 without the need to increase Zr doping level. Thus, compared to other approaches with high Zr doping (where x in Hf(1−x)ZrxO is greater than 0.7), Zr doping of less than 70% can be used for the ferroelectric structure 104 so dopants are more evenly distributed, and wake-up effect is reduced. Thereby, stability of the ferroelectric structure can be improved with better ferroelectric endurance.
Further still, in some examples, due to the presence of the interlayer 114, the ferroelectric layer 110 and/or anti-ferroelectric layer 112 can have a maximum grain size width and/or average grain size of more than 20 microns, which can be significantly larger than conventional grain sizes and can help to reduce current leakage along grain boundaries within ferroelectric structure 104. Referring briefly to
Although
A semiconductor device including multiple device terminals is disposed in and/or over the substrate. In the illustrated example, the semiconductor device is a MOSFET 305, but in other examples the semiconductor device could be another active device (e.g., bipolar junction transistor (BJT), FinFET, or junction FET (JFET) among others), a passive device (e.g., diode, resistor, or capacitor, among others). The illustrated MOSFET 305 includes a source region 306 and a drain region 308, both of which have a first conductivity type. A channel region 310 separates the source region 306 and drain region 308 and has a second conductivity type opposite the first conductivity type. A conductive gate electrode 312 is disposed over the channel region 310 and can comprise doped polysilicon or metal. A gate dielectric 314, which can be made of silicon dioxide or a high-k dielectric material, separates the gate electrode 312 from the channel region 310. Sidewall spacers 315, which can be made for example of silicon nitride, are disposed on outer edges of the gate electrode 312. A dielectric layer 316, such as silicon dioxide or a low-k dielectric is disposed over the gate electrode 312 and source region 306 and drain region 308. Contacts 318, which are made of a metal (e.g., tungsten, copper, aluminum, titanium, nickel, and/or tantalum), extend vertically from the source region 306, drain region 308 and gate electrode 312 though the dielectric layer 316 and through an etch stop layer 320.
The interconnect structure 304 include a first metal layer 322, second metal layer 324, and third metal layer 326, as well as higher metal layers (not shown). Each of these metal layers can include metal lines that extend into the plane of the page and/or across the page to connect semiconductor device (e.g., MOSFET 305) to one another, thereby achieving a desired circuit configuration. Vias (e.g., 328, 329) extend vertically through one or more of a plurality of stacked interlayer dielectric (ILD) layers 330a-330b and between adjacent metal layers to couple metal layers at different heights to one another. In some embodiments, the plurality of stacked ILD layers 330a-330b may comprise one or more of silicon dioxide, silicon nitride, carbon doped silicon dioxide, silicon oxynitride, borosilicate glass (BSG), phosphorus silicate glass (PSG), borophosphosilicate glass (BPSG), fluorosilicate glass (FSG), undoped silicate glass (USG), a porous dielectric material, or the like. In various embodiments, the one or more etch stop layers 332a-332b may comprise a carbide (e.g., silicon carbide, silicon oxycarbide, or the like), a nitride (e.g., silicon nitride, silicon oxynitride, or the like), or the like. The metal layers and vias typically comprise a metal, and may comprise copper, aluminum, tantalum, and/or titanium, among others.
In the example of
In
In the example of
The FeFET memory circuit 700 comprises a FeFET memory array 702 including a plurality of FeFET devices 7041,1-704n,m. The plurality of FeFET devices 7041,1-704n,m are arranged within the FeFET memory array 702 in rows and/or columns. The plurality of FeFET devices 7041,x-704n,x within a row are operably coupled to word-lines WLx (x=1-m). The plurality of FeFET devices 704x,1-704x,m have respective drains along a column that are operably coupled to bit-lines BLx (x=1-n) and have respective sources along a column that are operably coupled to source-lines SLx (x=1-n). Each FeFET device 704 can have the structure illustrated in any of
The word-lines WL1-WLm, the bit-lines BL1-BLn, and the source-lines SL1-SLn, are coupled to control circuitry 706. In some embodiments, the control circuitry 706 comprises a word-line decoder 710 coupled to the word-lines WL1-WLm, a bit-line decoder 708 coupled to the bit-lines BL1-BLn, and a source-line decoder 712 coupled to the source-lines SL1-SLn. In some embodiments, the control circuitry 706 further comprises a sense amplifier 714 coupled to the bit-lines BL1-BLn or the source-lines SL1-SLn. In some embodiments, the control circuitry 706 further comprises a control unit 716 configured to send address information SADR to the word-line decoder 710, the bit-line decoder 708, and/or the source-line decoder 412 to enable the control circuitry 706 to selectively access one or more of the plurality of FeFET devices 7041,1-704n,m.
For example, during operation, the control unit 716 is configured to provide address information SADR to the word-line decoder 710, the bit-line decoder 708, and the source-line decoder 712. Based on the address information SADR, the word-line decoder 710 is configured to selectively apply a bias voltage to one of the word-lines WL1-WLm. Concurrently, the bit-line decoder 708 is configured to selectively apply a bias voltage to one of the bit-lines BL1-BLn and/or the source-line decoder 712 is configured to selectively apply a bias voltage to one of the source-lines SL1-SLn. By applying bias voltages to selective ones of the word-lines WL1-WLm, the bit-lines BL1-BLn, and/or the source-lines SL1-SLn, the FeFET memory circuit 700 can be operated to write different data states to and/or read data states from the plurality of FeFET devices 7041,1-704n,m.
As shown in cross-sectional view 900 of
As shown in cross-sectional view 1000 of
As shown in cross-sectional view 1100 of
As shown in cross-sectional view 1200 of
In some embodiments, the first patterning process may selectively expose the oxide semiconductor layer and the ferroelectric layer to a first etchant 1202 according to a first masking structure 1204 formed over capping layers (e.g., 1104 of
As shown in cross-sectional view 1300 of
As shown in cross-sectional view 1400 of
As shown in cross-sectional view 1500 of
While the disclosed method 1600 is illustrated and described herein as a series of acts or events, it will be appreciated that the illustrated ordering of such acts or events are not to be interpreted in a limiting sense. For example, some acts may occur in different orders and/or concurrently with other acts or events apart from those illustrated and/or described herein. In addition, not all illustrated acts may be required to implement one or more aspects or embodiments of the description herein. Further, one or more of the acts depicted herein may be carried out in one or more separate acts and/or phases.
At act 1602, a ferroelectric layer is formed on a conductive gate layer.
At act 1604, an oxide semiconductor layer is formed on the ferroelectric layer, and a capping layer is formed over the oxide semiconductor layer.
At act 1606, one or more patterning processes are performed to remove portions of the capping layer and the ferroelectric structure to expose the oxide semiconductor.
At act 1608, a dielectric layer is formed over the ferroelectric structure and capping structure.
At act 1610, one or more additional patterning processes are performed to form a source contact hole and/or a drain contact hole that extend through the dielectric layer and though the capping structure to expose the oxide semiconductor.
At act 1612, a conductive material is formed within the source contact hole and/or the drain contact hole.
Accordingly, in some embodiments, a ferroelectric field-effect transistor (FeFET) device, includes a ferroelectric structure including a ferroelectric layer and an anti-ferroelectric layer; a gate structure disposed along a first face of the anti-ferroelectric layer, such that the anti-ferroelectric layer separates the gate structure from the ferroelectric layer; an oxide semiconductor disposed along a first face of the ferroelectric layer, such that the ferroelectric layer separates the oxide semiconductor from the anti-ferroelectric layer; a source region and a drain region disposed on the oxide semiconductor, wherein the gate structure extends laterally over the anti-ferroelectric layer between the source region and the drain region; and an interlayer separating the ferroelectric layer from the anti-ferroelectric layer.
In some examples, the ferroelectric layer comprises Hf1-xZrxO, and the anti-ferroelectric layer comprises Hf1-yZryO, where x and y are different.
In some examples, x is between 0 and 0.5, and y is between 0.5 and 1.
In some examples, the interlayer comprises a metal oxide or a metal.
In some examples, the interlayer comprises the metal oxide and the metal oxide comprises titanium oxide, aluminum oxide, magnesium oxide, hafnium oxide, or indium oxide; or wherein the interlayer comprises the metal and the metal comprises titanium, platinum, gold, or nickel.
In some examples, the ferroelectric layer has a first percentage of tetragonal crystalline lattice per unit volume and the anti-ferroelectric layer has a second percentage of tetragonal crystalline lattice per unit volume, the second percentage being greater than the first percentage.
In some examples, a ratio of a first percentage of orthorhombic crystal per unit volume in the ferroelectric layer to a second percentage of tetragonal crystal per unit volume in the anti-ferroelectric layer is in a range of 15:1 to 25:1.
In some examples, at least one of the ferroelectric layer or the anti-ferroelectric layer is a polycrystalline structure where crystalline grains are separated from one another by grain boundaries, and the crystalline grains have an average grain size of more than 20 microns.
In some examples, the interlayer has a coefficient of thermal expansion of 1×10-4 m/Kelvin (K) to 1×10-6 m/K.
In some examples, the interlayer has a thickness of less than 1 nanometer.
In some examples, the oxide semiconductor comprises one or more of indium gallium zinc oxide, indium gallium zinc tin oxide, indium tungsten oxide, indium tungsten zinc oxide, indium zinc oxide, or zinc oxide.
In some examples, the source region is coupled to a source-line, the drain region is coupled to a bit-line, and the gate structure is coupled to a word-line.
In some examples, FeFET device further includes a substrate; wherein the gate structure is disposed along a first side of the substrate; and wherein the gate structure is vertically disposed between the first side of the substrate and the ferroelectric structure.
In some examples, an integrated chip includes a semiconductor substrate; a conductive gate structure arranged over the semiconductor substrate; an anti-ferroelectric layer arranged over the conductive gate structure; an interlayer comprising a metal or metal oxide having a lower surface in direct contact with an upper surface of the anti-ferroelectric layer; a ferroelectric layer having a lower surface in direct contact with an upper surface of the interlayer; an oxide semiconductor arranged over the ferroelectric layer, the oxide semiconductor including a channel region directly over the conductive gate structure; and a source region and a drain region disposed on the oxide semiconductor and laterally spaced apart from one another by a length corresponding to the channel region.
In some examples, the ferroelectric layer has a first percentage of tetragonal crystalline lattice per unit volume and the anti-ferroelectric layer has a second percentage of tetragonal crystalline lattice per unit volume, the second percentage being greater than the first percentage.
In some examples, in the ferroelectric layer includes Hf1-xZrxO, and the anti-ferroelectric layer includes Hf1-yZryO, where x is less than 0.5 and greater than 0, and y is greater than 0.5 and less than 1.
In some examples, a method of forming a FeFET device includes receiving a substrate; forming a gate structure over the gate structure; forming an anti-ferroelectric layer over the gate structure; forming a interlayer comprising a metal or metal oxide having a lower surface in direct contact with an upper surface of the anti-ferroelectric layer; forming a ferroelectric layer having a lower surface in direct contact with an upper surface of the interlayer; forming an oxide semiconductor layer over the ferroelectric layer; forming a dielectric layer over the oxide semiconductor layer; performing a first patterning process to form a source opening and a drain opening through the dielectric layer to expose the oxide semiconductor layer; and forming a conductive material within the source opening and the drain opening.
In some examples, the anti-ferroelectric layer and the interlayer are formed in-situ within a deposition chamber.
In some examples, the anti-ferroelectric layer, the interlayer, and the ferroelectric layer are formed in-situ within a deposition chamber.
In some examples, the oxide semiconductor layer comprises indium gallium zinc oxide, indium gallium zinc tin oxide, indium tungsten oxide, indium tungsten zinc oxide, indium zinc oxide, or zinc oxide.
The foregoing outlines features of several embodiments so that those skilled in the art may better understand the aspects of the present disclosure. Those skilled in the art should appreciate that they may readily use the present disclosure as a basis for designing or modifying other processes and structures for carrying out the same purposes and/or achieving the same advantages of the embodiments introduced herein. Those skilled in the art should also realize that such equivalent constructions do not depart from the spirit and scope of the present disclosure, and that they may make various changes, substitutions, and alterations herein without departing from the spirit and scope of the present disclosure.