Ferroelectric HfO2 thin films 10 nm thick are simultaneously doped with Al and Si. The arrangement of the Al and Si dopant layers within the HfO2 greatly influences the resulting ferroelectric properties of the polycrystalline thin films. Optimizing the order of the Si and Al dopant layers led to a remanent polarization of ∼20 μC/cm2 and a coercive field strength of ∼1.2 MV/cm. Post-metallization anneal temperatures from 700 °C to 900 °C were used to crystallize the Al and Si doped HfO2 thin films. Grazing incidence x-ray diffraction detected differences in peak broadening between the mixed Al and Si doped HfO2 thin films, indicating that strain may influence the formation of the ferroelectric phase with variations in the dopant layering. Endurance characteristics show that the mixed Al and Si doped HfO2 thin films exhibit a remanent polarization greater than 15 μC/cm2 up to 108 cycles.

The emergence of ferroelectricity in HfO2-based thin films is remarkable due to the advancements it may enable for developing and emerging memory technologies such as ferroelectric random access memory (FRAM) and ferroelectric field effect transistors (FeFETs).1 Ferroelectric behavior in HfO2-based thin films can be generated through a suitable combination of doping and annealing conditions.2 Various dopants have promoted ferroelectricity in HfO2 thin films including Si, Al, Y, Gd, Sr, La, and Zr.2–9 Atomic layer deposition (ALD) is most frequently used to deposit HfO2 thin films because of its ability to tailor precise doping profiles10 and its highly conformal growth process which can enable high aspect ratio device structures.11 Ferroelectric HfO2-based thin films are typically grown in an amorphous state and then crystallized during a rapid thermal anneal (RTA) with a capping electrode. A capping electrode during rapid thermal annealing enhances and often facilitates ferroelectric behavior in HfO2-based thin films.3–5 

The combination of chemical doping and a mechanically constraining top electrode is postulated to encourage a tetragonal to orthorhombic (t → o) phase transformation during the crystallization anneal.2 A noncentrosymmetric, polar Pca21 orthorhombic phase is the candidate source for ferroelectricity in HfO2 thin films. Atomic imaging has provided some structural evidence for the existence of this crystal phase,12 although a mixture of phases may be present in polycrystalline HfO2 thin films after rapid thermal annealing.13 The coexistence of the monoclinic, tetragonal, orthorhombic, and cubic phases may thus become manifest in doped polycrystalline HfO2 thin films. In order to maximize the remanent polarization (Pr) of ferroelectric HfO2, the fraction of the polar orthorhombic crystal phase should be increased within the thin films to enlarge the magnitude of switched charge.

While ferroelectric Hf0.5Zr0.5O2 is attractive due its simple binary composition and low thermal budget,13 more flexibility in engineering the properties of HfO2-based ferroelectrics may be achieved with other dopants. For instance, adjusting the Si-dopant layering distribution in HfO2 thin films integrated into the metal-ferroelectric-insulator-semiconductor (MFIS) structure illustrated how the Pr could be lowered by increasing the distance between the Si layers.10 FeFETs may benefit from the engineering of the ferroelectric for lower Pr's because of the consequent reduction in the depolarization field.14 Furthermore, the small quantity of Al and Si dopants required to induce ferroelectricity in HfO2 thin films opens up the possibility of placing the dopants within the HfO2 stack simultaneously. In this letter, we show that mixed Al and Si-doped HfO2 thin films exhibit a wide range of ferroelectric properties through the placement of the dopant layers within the HfO2 film.

ALD TiN bottom electrodes 10 nm thick were deposited at 200 °C on highly doped (0.001–0.005 Ω cm) (100) p + Si wafers. Tetrakis(dimethylamido)titanium was used as the Ti precursor and exposed to N2 plasma during TiN growth. HfO2 thin films doped with both Al and Si were grown by thermal ALD at 200 °C. The Hf, Al, and Si precursors used were tetrakis(dimethylamido)hafnium, trimethylaluminum (TMA), and tris(dimethylamino)silane, respectively. A 24:1 Hf:(Al, Si) pulse ratio was maintained where a total of 4 dopant layers were incorporated into the 10 nm thick HfO2 thin films. Three different distributions of the Al and Si dopant layers within the film stack were fabricated. The three different dopant distributions may be denoted as Si-Al-Al-Si (0.87 mol. % Al, 0.49 mol. % Si), Al-Si-Si-Al (1.08 mol. % Al, 0.75 mol. % Si), and Al-Si-Al-Si (1.12 mol. % Al, 0.68 mol. % Si), which represent the placement of the dopant layers from the bottom-up within the HfO2 film. Time of flight secondary ion mass spectrometry (TOF-SIMS) was employed to quantify the Al and Si content within the HfO2 thin films. Approximately 2.4 nm of HfO2 are in between each dopant layer in the as-grown films. ALD TiN top electrodes 10 nm thick were deposited at 200 °C. In previous work, we had observed ferroelectric behavior in Si-doped HfO2 with a 24:1 HfO2:SiO2 ALD cycle ratio.15 Thus, the 24:1 ALD cycle ratio was chosen to investigate mixed Al and Si doping in HfO2 thin films. The entire TiN metal-ferroelectric-metal (MFM) stack was grown within the ALD chamber at 200 °C to prevent oxidation of the HfO2/electrode interface.15 The MFM stacks were annealed at 700 °C–900 °C for 20 s in N2. 50 nm thick Pt contacts were deposited in a liftoff process and used as a hard mask in an SC1 wet etch for capacitor geometry formation. The capacitor area was 3600 μm2.

The hysteresis characteristics, also referred to as polarization vs. electric field (P–E) plots, of the mixed Al and Si doped HfO2 thin films were extracted from a Sawyer-Tower circuit with a 220 nF linear capacitor and a 1 kHz triangle wave. Virgin device characteristics were extracted from the average of the first 50 hysteresis measurements. Endurance cycling was performed at 1 MHz with bipolar square waves and a 50 Ω resistor in series with the ferroelectric capacitor during cycling. Grazing incidence x-ray diffraction (GIXRD) was performed with a Rigaku SmartLab X-Ray Diffractometer with a grazing incidence angle of 0.5°, a 0.05° step size, and a 4 s count time per step.

Simultaneously doping HfO2 thin films with spatially segregated Al and Si dopant layers may influence the internal strain of the ferroelectric. The strain in ferroelectric HfO2 impacts the formation of the ferroelectric o-phase16 and could alter the magnitude of the coercive field.17 Internal strain within HfO2 can be expected to arise from mixed doping in view of the experimentally extracted differences in the HfO2 lattice volume between pure Al (127.88 Å3) and Si (129.57 Å3) doped HfO2 thin films.17 In-plane tensile stress is thought to promote the favorable t → o phase transformation16 and enhance the Pr for ferroelectric HfO2 thin films.

The hysteresis characteristics of the mixed Al and Si doped HfO2 thin films reveal significant differences in the Pr and overall shape of the P-E loop before and after electric field cycling at anneal temperatures of 800 °C and 900 °C, Figs. 1(a)–1(d). The virgin characteristics of the Si-Al-Al-Si doped HfO2 thin films exhibit a Pr of 19 μC/cm2 which increases by ∼1 μC/cm2 after cycling. The P-E loops of both the Al-Si-Si-Al and the Al-Si-Al-Si doped HfO2 thin films look distorted horizontally in their virgin state with a Pr of 13 μC/cm2 and 7 μC/cm2, respectively. Cycling subsequently increases the Pr in the two films by 2 μC/cm2 and 5 μC/cm2, respectively. Electric field cycling, sometimes referred to as the “wake-up” effect, has been postulated to increase Pr and decrease the asymmetry in ferroelectric Si-doped HfO2 through the redistribution of mobile defect charge (i.e., oxygen vacancies) originating from the electrode interfaces.15,18 Since the mixed Al and Si doped HfO2 thin films show unequal enhancements of Pr with electric field cycling, lattice relaxations associated with the promotion of the o-phase may occur at different rates within the differently doped films. At 900 °C, the Pr trends amongst the three doping layering distributions remain similar, although a significant increase in the coercive field occurs for all of the films. The remanent polarization value of ∼20 μC/cm2 found in these mixed Al and Si doped HfO2 thin films are in good agreement with theoretical simulations of the spontaneous polarization predicted for the Pca21 phase of ∼52 μC/cm2, which is expected to be double the Pr found experimentally.17,19 The mixed Al and Si doped HfO2 thin films exhibit remanent polarization values comparable to Si-, Al-, and Y-doped HfO2, although higher Pr values have been reported for Gd and La-doped HfO2 thin films.8 

FIG. 1.

Hysteresis characteristics of the mixed Al and Si doped HfO2 thin films annealed at 800 °C (a) before and (b) after 5000 cycles. Annealing the films at 900 °C for 20 s led to increases in the coercive field in the P-E of (c) the virgin and (d) cycled capacitors. Cycling was performed with ±3.5 MV/cm bipolar square waves at 1 kHz.

FIG. 1.

Hysteresis characteristics of the mixed Al and Si doped HfO2 thin films annealed at 800 °C (a) before and (b) after 5000 cycles. Annealing the films at 900 °C for 20 s led to increases in the coercive field in the P-E of (c) the virgin and (d) cycled capacitors. Cycling was performed with ±3.5 MV/cm bipolar square waves at 1 kHz.

Close modal

GIXRD was used to confirm that all of the mixed Al and Si doped HfO2 thin films show the possible existence of the Pca21 orthorhombic phase, Fig. 2. GIXRD indicates the coexistence of multiple phases within the doped HfO2 thin films. The full width at half maximum (FWHM) of the t(111)/o(211) diffraction peak around 30° is largest in the Si-Al-Al-Si and smallest in the Al-Si-Al-Si doped HfO2 thin films. Peak width broadening of the mixed doped HfO2 thin films supports the hypothesis that the doping distribution of the Al and Si layers may be influencing the strain of the HfO2 thin film, although changes in grain size cannot be ruled out. Since the largest peak width broadening was found in the Si-Al-Al-Si doped HfO2 thin films, these films may have the greatest amount of microstrain. The sharp reflections in the GIXRD plots at ∼52° are attributed to the signal from the single crystal quartz sample holder.

FIG. 2.

GIXRD confirms the possible existence of the orthorhombic Pca21 phase in all three of the Al and Si doped HfO2 thin films annealed at 800 °C.

FIG. 2.

GIXRD confirms the possible existence of the orthorhombic Pca21 phase in all three of the Al and Si doped HfO2 thin films annealed at 800 °C.

Close modal

The way in which the different dopants are layered within the HfO2 thin films affects the formation of the orthorhombic phase through the induced strain and competing phase transformations. Since the Si-Al-Al-Si doped HfO2 thin films exhibit the highest remanent polarization and the most significant peak broadening, it is likely that strain is aiding in the stabilization of the orthorhombic phase. Simulations have shown that the tetragonal P42/nmc phase may transition into the polar orthorhombic Pca21 through a suitable combination of stress and/or electric field.19 In particular, the preferential formation of the o-phase in the Si-Al-Al-Si distribution may be generated through a large in-plane tensile strain due to the smaller Al-doped HfO2 lattice volume compared to the larger Si-doped HfO2 lattice volume.16,17 The doping distribution of Al and Si in the annealed HfO2 thin films was confirmed with TOF-SIMS, Fig. 3. The Gaussian shape of the Si and Al distributions is caused by the limited depth resolution of TOF-SIMS. The dopant layers are consistent with a monolayer or sub-monolayer of Al or Si.

FIG. 3.

TOF-SIMS depth profiles of the (a) Al-Si-Al-Si, (b) Al-Si-Si-Al, and (c) Si-Al-Al-Si doped HfO2 thin films annealed at 800 °C.

FIG. 3.

TOF-SIMS depth profiles of the (a) Al-Si-Al-Si, (b) Al-Si-Si-Al, and (c) Si-Al-Al-Si doped HfO2 thin films annealed at 800 °C.

Close modal

High resolution cross-sectional transmission electron microscopy (HR-TEM) confirmed that the Al and Si doped HfO2 thin films were crystallized after annealing, Fig. 4. The Pr dependence on annealing temperature and electric field cycling are shown for the mixed Al and Si-doped HfO2 thin films in Figs. 5(a)–5(d). The Al-Si-Al-Si doped HfO2 thin films favor a lower anneal temperature and reach a maximum virgin Pr of ∼9 μC/cm2 at 700 °C. The Al-Si-Si-Al doped HfO2 thin films reach a maximum virgin Pr of ∼14 μC/cm2 at 850 °C. An anneal temperature range from 800 °C to 900 °C yields the highest Pr values on the order of ∼20 μC/cm2 in the Si-Al-Al-Si doped HfO2 thin films with the “wake-up” effect being noticeably absent for the 800 °C and 900 °C anneals. These trends indicate that the proper selection of annealing conditions to maximize Pr is highly dependent on the doping distributions employed in HfO2-based thin films. Endurance cycling characteristics of the Si-Al-Al-Si doped HfO2 thin films annealed at 800 °C for 20 s are shown in Fig. 5(d). At 3.5 MV/cm, a Pr greater than 15 μC/cm2 is observed up to 108 until breakdown. Breakdown is accelerated at 4 MV/cm and occurs after 106 cycles. Cycling the mixed Al and Si doped HfO2 thin film at 3 MV/cm delays breakdown until 109 cycles at the cost of a lower Pr and the acceleration of fatigue. The cycling characteristics are promising for memory applications such as FRAM, which require a very high number of read and write cycles. Further optimization of the processing conditions, including the doping and annealing conditions, is likely to provide greater improvements in the maximum number of cycles HfO2-based ferroelectrics can withstand before breakdown.

FIG. 4.

HR-TEM image of the Si-Al-Al-Si doped HfO2 thin film after annealing at 850 °C. Lattice fringes can be seen in the polycrystalline TiN electrodes.

FIG. 4.

HR-TEM image of the Si-Al-Al-Si doped HfO2 thin film after annealing at 850 °C. Lattice fringes can be seen in the polycrystalline TiN electrodes.

Close modal
FIG. 5.

Anneal temperature dependence of the (a) Al-Si-Al-Si, (b) Al-Si-Si-Al, and (c) Si-Al-Al-Si doped HfO2 thin films. The black dots denote the virgin Pr and the red squares represent the Pr after 5000 cycles at 1 kHz, 3.5 MV/cm. (d) Endurance cycling of the Si-Al-Al-Si doped HfO2 thin films annealed at 800 °C for different applied electric fields.

FIG. 5.

Anneal temperature dependence of the (a) Al-Si-Al-Si, (b) Al-Si-Si-Al, and (c) Si-Al-Al-Si doped HfO2 thin films. The black dots denote the virgin Pr and the red squares represent the Pr after 5000 cycles at 1 kHz, 3.5 MV/cm. (d) Endurance cycling of the Si-Al-Al-Si doped HfO2 thin films annealed at 800 °C for different applied electric fields.

Close modal

In summary, HfO2 thin films 10 nm thick were simultaneously doped with Al and Si with different dopant layering distributions. The resulting ferroelectric properties in the mixed Al and Si doped HfO2 thin films were dependent on the layering distribution and the anneal temperature. The influence of strain on the formation of the ferroelectric o-phase may be supported by differences in peak broadening measured from GIXRD, which was found to be dependent on the Al and Si doping distribution. The Al and Si doped HfO2 thin films were cycled up to 108 times with an applied electric field strength of 3.5 MV/cm before breakdown occurred. The ability to engineer the ferroelectric properties of HfO2 thin films by simultaneously doping HfO2 thin films with multiple dopant species is thus presented as an accessible route for enhancing HfO2-based ferroelectrics for a variety of applications.

This work was supported in part by the NSF I/UCRC on Multi-functional Integrated System Technology (MIST Center) IIP-1439644. P.L. is the recipient of a SRC Graduate Fellowship, and J.J. acknowledges support from NSF under DMR-1207293 (J.J., C.Z.). The authors acknowledge the use of the Analytical Instrumentation Facility (AIF) at North Carolina State University, which was supported by the State of North Carolina and the National Science Foundation.

1.
J.
Müller
,
P.
Polakowski
,
S.
Riedel
,
S.
Mueller
,
E.
Yurchuk
, and
T.
Mikolajick
, in
2014 14th Annual Non-Volatile Memory Technologies Symposium (NVMTS)
(
2014
), pp.
1
7
.
2.
T. S.
Böscke
,
J.
Müller
,
D.
Braühaus
,
U.
Schröder
, and
U.
Böttger
,
Appl. Phys. Lett.
99
,
102903
(
2011
).
3.
P. D.
Lomenzo
,
P.
Zhao
,
Q.
Takmeel
,
S.
Moghaddam
,
T.
Nishida
,
M.
Nelson
,
C. M.
Fancher
,
E. D.
Grimley
,
X.
Sang
,
J. M.
LeBeau
, and
J. L.
Jones
,
J. Vac. Sci. Technol., B: Microelectron. Nanometer Struct. Process., Meas., Phenom.
32
,
03D123
(
2014
).
4.
S.
Mueller
,
J.
Mueller
,
A.
Singh
,
S.
Riedel
,
J.
Sundqvist
,
U.
Schroeder
, and
T.
Mikolajick
,
Adv. Funct. Mater.
22
,
2412
(
2012
).
5.
J.
Müller
,
U.
Schröder
,
T. S.
Böscke
,
I.
Müller
,
U.
Böttger
,
L.
Wilde
,
J.
Sundqvist
,
M.
Lemberger
,
P.
Kücher
,
T.
Mikolajick
, and
L.
Frey
,
J. Appl. Phys.
110
,
114113
(
2011
).
6.
S.
Mueller
,
C.
Adelmann
,
A.
Singh
,
S.
Van Elshocht
,
U.
Schroeder
, and
T.
Mikolajick
,
ECS J. Solid State Sci. Technol.
1
,
N123
(
2012
).
7.
T.
Schenk
,
S.
Mueller
,
U.
Schroeder
,
R.
Materlik
,
A.
Kersch
,
M.
Popovici
,
C.
Adelmann
,
S.
Van Elshocht
, and
T.
Mikolajick
, in
2013 Proceedings of European Solid-State Device Research Conference
(
IEEE
,
2013
), pp.
260
263
.
8.
U.
Schroeder
,
E.
Yurchuk
,
J.
Müller
,
D.
Martin
,
T.
Schenk
,
P.
Polakowski
,
C.
Adelmann
,
M. I.
Popovici
,
S. V.
Kalinin
, and
T.
Mikolajick
,
Jpn. J. Appl. Phys., Part 1
53
,
08LE02
(
2014
).
9.
J.
Müller
,
T. S.
Böscke
,
D.
Bräuhaus
,
U.
Schröder
,
U.
Böttger
,
J.
Sundqvist
,
P.
Kücher
,
T.
Mikolajick
, and
L.
Frey
,
Appl. Phys. Lett.
99
,
112901
(
2011
).
10.
P. D.
Lomenzo
,
Q.
Takmeel
,
C.
Zhou
,
Y.
Liu
,
C. M.
Fancher
,
J. L.
Jones
,
S.
Moghaddam
, and
T.
Nishida
,
Appl. Phys. Lett.
105
,
072906
(
2014
).
11.
P.
Polakowski
,
S.
Riedel
,
W.
Weinreich
,
M.
Rudolf
,
J.
Sundqvist
,
K.
Seidel
, and
J.
Müller
, in
2014 IEEE 6th International Memory Workshop
(
IEEE
,
2014
), pp.
1
4
.
12.
X.
Sang
,
E. D.
Grimley
,
T.
Schenk
,
U.
Schroeder
, and
J. M.
LeBeau
,
Appl. Phys. Lett.
106
,
162905
(
2015
).
13.
M. H.
Park
,
H. J.
Kim
,
Y. J.
Kim
,
W.
Lee
,
T.
Moon
, and
C. S.
Hwang
,
Appl. Phys. Lett.
102
,
242905
(
2013
).
14.
T. P.
Ma
,
IEEE Electron Device Lett.
23
,
386
(
2002
).
15.
P. D.
Lomenzo
,
Q.
Takmeel
,
C.
Zhou
,
C. M.
Fancher
,
E.
Lambers
,
N. G.
Rudawski
,
J. L.
Jones
,
S.
Moghaddam
, and
T.
Nishida
,
J. Appl. Phys.
117
,
134105
(
2015
).
16.
M. H.
Park
,
H. J.
Kim
,
Y. J.
Kim
,
T.
Moon
, and
C. S.
Hwang
,
Appl. Phys. Lett.
104
,
072901
(
2014
).
17.
S.
Clima
,
D. J.
Wouters
,
C.
Adelmann
,
T.
Schenk
,
U.
Schroeder
,
M.
Jurczak
, and
G.
Pourtois
,
Appl. Phys. Lett.
104
,
092906
(
2014
).
18.
D.
Zhou
,
J.
Xu
,
Q.
Li
,
Y.
Guan
,
F.
Cao
,
X.
Dong
,
J.
Müller
,
T.
Schenk
, and
U.
Schröder
,
Appl. Phys. Lett.
103
,
192904
(
2013
).
19.
T. D.
Huan
,
V.
Sharma
,
G. A.
Rossetti
, and
R.
Ramprasad
,
Phys. Rev. B
90
,
064111
(
2014
).