Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2024 Jan 18;6(2):770–776. doi: 10.1021/acsaelm.3c01231

Dependence of the Metal–Insulator–Semiconductor Schottky Barrier Height on Insulator Composition

Benjamin E Davis 1,*, Nicholas C Strandwitz 1
PMCID: PMC10902844  PMID: 38435804

Abstract

graphic file with name el3c01231_0005.jpg

The effects of different high-κ tunnel oxides on the metal–insulator–semiconductor Schottky barrier height (ΦB) were systematically investigated. While these high-κ interlayers have been previously observed to affect ΦB, there has never been a clear consensus as to why this ΦB modulation occurs. Changes in ΦB were measured when adding 0.5 nm of seven different high-κ oxides to n-Si/Ni contacts with a thin native silicon oxide also present. Depending on the high-κ oxide composition and ΦB measurement technique, increases in ΦB up to 0.4 eV and decreases up to 0.2 eV with a high-κ introduction were measured. The results were compared to several different hypotheses regarding the effects of tunnel oxides on ΦB. The experimental data correlated most closely with the model of a dipole formed at the SiOx/high-κ interface due to the difference in the oxygen areal density between the two oxides. Knowledge of this relationship will aid in the design of Schottky and ohmic contacts by providing criteria to predict the effects of different oxide stacks on ΦB.

Keywords: Schottky barriers, atomic layer deposition, tunnel oxides, thin films, interface dipoles

Introduction

It is desirable to control the Schottky barrier height (ΦB) at the metal–semiconductor contacts for a variety of applications. For example, ΦB minimization is needed for ohmic contacts. Decreasing ΦB is a key factor in reducing parasitic resistance at transistor source/drain contacts, which is of importance for continued scaling down of device dimensions.1−3 The formation of a Schottky barrier at electrical contact interfaces also impedes charge carrier collection in certain photovoltaic architectures.4,5 In other cases, maximization of ΦB is desirable. In Schottky barrier solar cells, a large ΦB is necessary to separate photogenerated electron–hole pairs and prevent recombination.6,7 Power rectifiers utilize a large ΦB to achieve a high turn-on voltage and reduce leakage current in the “off” state.8 Variation in interfaces and insulating layers in metal–insulator–semiconductor (MIS) stacks has a dramatic impact on ΦB but an accurate predictive framework for these systems does not yet exist.

According to the basic Schottky model, the barrier height between a metal and an n-type semiconductor is only a function of the vacuum metal work function ΦM,vac and the semiconductor electron affinity ΧS9

graphic file with name el3c01231_m001.jpg 1

In practice, eq 1 rarely agrees with experimental values of ΦB. Often, the metal work function has little effect on the value of ΦB. According to the concept of Fermi level pinning

graphic file with name el3c01231_m002.jpg 2

where ΦCNL is the charge neutrality level in the semiconductor and S is a “pinning” factor that determines the sensitivity of ΦB to ΦM,vac. Thus, eq 2 utilizes an “effective” metal work function that depends on the pinning factor and the charge neutrality level, instead of solely ΦM,vac.10

If an electrostatic dipole moment (ΔΦB,d) is also present at the contact, then it is also expected to impact ΦB

graphic file with name el3c01231_m003.jpg 3

Insertion of a tunnel active insulating layer between the metal and semiconductor can affect ΦB through changes in the S factor and interface moment. Several possible reasons for the resulting effect on ΦB have been discussed in the literature, including the insulator causing “depinning” of the Fermi level (i.e., a higher value of S and greater influence of the metal work function),11 the electrostatic effect of dielectric fixed charges, and the potential to form electric dipoles at interfaces. Aluminum oxide fixed charges may contribute to changes in ΦB on GaAs substrates.12 However, experiments have also shown that the alumina fixed charge is not a dominating factor of ΦB on Si substrates.13,14 Shifts in the flat band voltage of MIS capacitor structures have been observed when stacking high-κ gate oxides onto interfacial SiOx, and these shifts are often attributed to dipoles forming at the SiOx/high-κ interface.15−17 It was also experimentally demonstrated that such dipoles can be used to tune ΦB of MIS tunnel structures.3,13,18

Several explanations of the origin of the dipoles have been proposed. Kita and Toriumi suggested that dipoles form due to differences in the oxygen areal density (OAD) between two oxides.19 In this case, bond relaxation would be achieved when negatively charged oxygen ions move from the higher-OAD to the lower-OAD side of the interface, forming Frenkel-type defects and inducing a dipole with a positive charge in the high-OAD side and negative charge in the low-OAD side. This hypothesis is supported by a report on the correlation of XPS-measured electrical dipoles at SiOx/high-κ interfaces to the XPS-derived oxygen density ratios of the oxides.20

Calculations by Lin and Robertson instead indicated that dipoles originate from the parent metal work functions/group electronegativities as long as there was a sufficient difference in the dielectric constant κ between the layers.21 They argue that dipoles exist along the metal–oxygen bonds on both sides of the oxide interface with dipole strengths related to the parent metal work functions. The relative dielectric constants of the oxide layers then determine to what extent each dipole is “screened” and whether a net dipole will form. Zhu et al. measured SiOx/high-κ interfacial dipoles via photoelectron spectroscopy and observed that the dipole strength was correlated with the thickness of an interfacial silicate layer between SiOx and AlOx.22

Several experimental studies have attributed changes in the MIS ΦB to changes in dipoles formed between insulating layers.3,13,18,23 However, such electrical experiments on Si have only been performed with a limited selection of high-κ tunnel oxide compositions (namely, AlOx and LaOx). An experimental study on a broader selection of oxides provides a platform to test the above hypotheses. Thus, the present study measures the changes in ΦB upon the introduction of a series of insulator compositions in Si-based MIS structures and compares the results to the hypotheses discussed above.

Experimental Section

For diode fabrication (Figure S1), n-type Si substrates with doping density 9 × 1016 cm–3 were cleaned with UV–ozone (Jelight) for 5 min and immersed in 5 wt % hydrofluoric acid for 1 min. Immediately after the etch, substrates were sonicated for 5 min in isopropanol (IPA). Finally, substrates were cleaned using the standard RCA process that included sequential immersion in 1:1:5 (by volume) NH4OH/H2O2/H2O and 1:1:6 HCl/H2O2/H2O solutions for 10 min each at temperatures between 70 and 80 °C.24 After each wet cleaning step, samples were rinsed with 18.2 MΩ water.

Each sample set contained two control samples with only the terminal RCA SiOx. Atomic layer deposition (ALD) was used to grow the deposited oxide on other samples using a Cambridge Nanotech Savannah S100 reactor. The deposition parameters for each oxide are shown in Table 1.

Table 1. ALD Parameters for Growth of Each Oxide.

oxide metal precursor precursor temp. (°C) N2 boost? (Y/N) O source reactor temp. (°C) growth rate (Å/cycle) number of cycles
AlOx trimethylaluminum ∼30 N H2O 150 0.90 6
HfOx tetrakis(dimethylamido) hafnium 75 N H2O 150 1.1 5
TiOx tetrakis(dimethylamido) titanium 75 N H2O 150 0.57 9
MgOx bis(ethylcyclopentadienyl) magnesium 90 N H2O 150 1.3 4
NbOx (tert-butylimido)tris(diethylamino) niobium 90 Y H2O 150 0.54 9
SrOx bis(n,N′-di-tert-butylacetamidinato) strontium(II)dimer 90 Y O3 150 0.84 6
LaOx tris(N,N′-di-i-propylformamidinato) lanthanum 160 Y O3 200 1.4 4
Open in a new tab

After ALD, ∼200 nm thick circular Ni top contacts (area = 0.005 cm2) were deposited by electron-beam evaporation through a shadow mask. Back contacts were made by scratching through the chemical oxide on the back of each sample with a diamond tip, applying Ag paste to the scratched area, and using the paste to fix an Al plate to the sample.

Metal–oxide–semiconductor capacitors (MOSCAPs) were prepared for each deposited high-κ oxide. The MOSCAPs were prepared identically to the diodes except for two details: the samples were etched in HF a second time immediately prior to ALD to minimize interfacial SiOx, and 10 nm of oxide was deposited instead of 0.5 nm.

ΦB was measured from diodes using Mott–Schottky (MS) and current–voltage–temperature (I–V–T) methods. For MS measurements, capacitance–voltage (C–V) data were collected at 1 MHz using a Hewlett-Packard 4194a Impedance/Gain-Phase Analyzer.25 ΦB for each condition was averaged over 18 contact pads across two samples. For I–V–T measurements, I–V data were collected at a variety of temperatures using a Keithley 2450 Sourcemeter.26 The reverse bias current data were used for the fit. C–V measurements were also performed at 1 MHz on the MOSCAPs for the evaluation of κ, the flat-band voltage Vfb, and the density of interface states Dit via the Terman method for each oxide.27

Ten nm high-κ oxide films were measured by X-ray photoelectron spectroscopy (XPS) in a SPECS PHOIBOS 150 NAP-XPS system under ultra-high vacuum to probe the metal oxidation state(s) and oxygen/metal atomic ratio. High-sensitivity low-energy ion scattering was performed with an ION-TOF Qtac100 instrument to investigate substrate coverage after the growth of nominally 0.5 nm of each high-κ film. Finally, the densities of some of the films were measured using X-ray reflectivity in a Panalytical Empyrean diffractometer. Film thicknesses were measured using a J.A. Woollam VASE Spectroscopic Ellipsometer.

In an effort to test trends in ΦB against the theories discussed above, it was necessary to estimate the OAD of each oxide (Table 2). The OAD was estimated by the method described by Kita and Toriumi.19 This method expresses OAD σ = Vu–2/3, where Vu is the volume of a unit cell containing a single oxygen atom (e.g., Al2/3O or Hf1/2O), equal to the formula weight in grams divided by the film density in g/cm3. The film density was taken from the literature only when data were available from films synthesized using the same ALD precursors and deposition temperature employed in the present study.

Table 2. Parameters Used to Estimate OAD and Obtained Values for Each Oxide.

oxide formula film density (g/cm3) density source OAD (cm–2)
Al2O3 2.9 various methods28 1.48 × 1015
HfO2 9.4 XRR29 1.42 × 1015
TiO2 3.7 XRR30 1.45 × 1015
MgO 3.4 ± 0.2 XRR (in-house) 1.37 × 1015
Nb2O5 3.8 ± 0.4 XRR (in-house) 1.23 × 1015
SrO 3.1 ± 0.1 XRR (in-house) 6.90 × 1014
La2O3 6.2 ± 0.1 XRR (in-house) 1.07 × 1015
Open in a new tab

Results and Discussion

Nonlinear 1/C2 vs V curves were observed, so data were corrected by subtracting the “excess capacitance” resulting in more linear 1/C2 vs V data and correspondingly nearly constant ΦB with applied bias (Figure S2).31 Excess capacitance has been observed as the result of charging and discharging of trap states, and thus may be due to the lack of postdeposition anneal step in this work.9

The ΦB values of the MIS Schottky diodes with native SiOx or native SiOx and an additional high-κ oxide were quantified using C–V (Figure 1a) and I–V–T (Figure 1b) techniques. Without a deposited oxide (only native SiOx), ΦB values were ∼0.5 eV. Some deviation of ΦB was found among the SiOx control samples and, hence, the “shifts” in ΦB within each sample set, rather than the raw values of ΦB, are considered below. The insertion of a high-κ oxide increased (AlOx, HfOx, MgOx, NbOx, and TiOx) or decreased (SrOx, LaOx), depending on oxide composition. The magnitudes of the increases with the insertion of the high-κ layer were approximately 0.3–0.4 eV and the magnitude of the decreases was ∼0.1 eV for the C–V–determined ΦB. Qualitatively, I–V–T-determined trends were the same, but overall ΦB values were smaller. This behavior is consistent with ΦB inhomogeneities across the contacts, as measured current is dominated by lower-ΦB “patches” within the interface, whereas C–V measurements yield the mean ΦB across the interface area.32,33 However, the agreement in the trends between the two measurement techniques allows for generalization about the impact of the high-κ composition on ΦB. Limited data exist in the literature on ΦB modulation utilizing high-κ layers as thin as those in the present study. However, the I–V–T-measured ΦB increase of 0.1 eV with AlOx addition is consistent with a corresponding 0.1 eV ΦB decrease measured on p-type Si in ref (18), which used a similar ALD process to deposit AlOx onto chemical SiOx.

Figure 1.

Figure 1

Open in a new tab

Schottky barrier heights measured via (a) the C–V method and (b) the I–V–T method. The error bars in (a) represent the standard deviations. The inset of (b) is a schematic of the test structure used.

Single temperature current density versus voltage (J–V) measurements also provide a qualitative assessment of ΦB (Figure 2). Specifically, if a similar tunnel barrier is presented to the majority carrier electrons by each oxide layer, and assuming recombination currents are small relative to thermionic emission currents, the reverse saturation current densities (Jsat, V < 0 V) should be inversely related to ΦB. Indeed, the LaOx and SrOx samples exhibited larger Jsat values in reverse bias relative to the SiOx control sample, and the diodes with the other oxide compositions displayed lower Jsat values relative to a representative SiOx control sample. Further, the diodes with oxides that yielded the largest ΦB values in MS measurements displayed the lowest Jsat values (i.e., AlOx and HfOx). Quantitatively, the log of the saturation current density plotted against the C–V and I–V–T-derived ΦB exhibits linear R2 values of 0.83 and 0.34, respectively (Figure S3). A poor fit is not surprising as the current values are impacted not only by ΦB but also by the oxide tunnel barrier, which will be affected by oxide composition. However, the qualitative trend provides further evidence of the correct quantification of ΦB values and trends in these samples. Further, changes in Jsat values over 2 orders of magnitude with the same semiconductor substrate and metal composition demonstrate the practical importance of ΦB control with interfacial layers.

Figure 2.

Figure 2

Open in a new tab

Representative J–V data were collected from diodes with each type of tunnel oxide insulator at room temperature.

LEIS measurements were used to examine whether full coverage of the Si substrate was achieved by the high-κ oxide depositions or whether films contained exposed SiOx surfaces (Figure 3). The ALD films were compared with a Si substrate coated in thermally grown SiO2 that was cleaned with the standard RCA process, treated with atomic oxygen, and sputtered using a 0.5 keV Ar + ion beam with a fluence of 1 × 1015 ions/cm2. LEIS is highly sensitive to ions scattered by the outermost layer of atoms, in which case a “surface peak” is observed as can be seen for Si and Mg in the thermal SiO2 and MgOx samples, respectively. Subsurface scattering from a given element is also observed and will result in ion detection at lower energies than this surface peak.34 Thus, the absence of a surface peak for Si is expected to be a good qualitative criterion for the total coverage of the SiOx surface by the high-κ film. AlOx was not included in these measurements because the Al and Si peaks are too close together to be easily distinguished. The NbOx spectrum contains a peak between 1700 and 1750 eV, where the Si peak is observed, indicating that some Si surface atoms are exposed and the NbOx coverage is not complete. None of the other high-κ oxides exhibited a surface peak at these energies, indicating that the surface was completely covered with the deposited high-κ oxide. All oxides exhibited increases in ion yield at ion energies lower than 1800 eV due to subsurface scattering from Si below the deposited oxide surface. Regardless of incomplete coverage, NbOx had a demonstrable effect on ΦB and was still considered in the overall trends below.

Figure 3.

Figure 3

Open in a new tab

LEIS spectra in the Si energy range from 0.5 nm films of each high-κ oxide and a thermal SiO2 substrate (top).

Several potential factors that may influence ΦB are compared against the experimental ΦB data including the parent metal work function, parent metal Pauling electronegativity, measured oxide dielectric constant, and estimated OAD (Figure 4). These comparisons are made based on the C–V data; however, similar results were obtained fitting the qualitatively similar I–V–T data. A weak positive association was found between the observed ΦB shifts and the known parent metal work functions (R2 = 0.62, Figure 4a), as well as the parent metal Pauling electronegativities (R2 = 0.56, Figure 4b).35 NbOx was excluded from work function analysis because of the wide range of work functions reported for Nb.35 No apparent trend was found between the ΦB shifts and the measured dielectric constants of the ALD oxides (R2 = 0.09, Figure 4c). However, when comparing the shifts with the estimated OAD values, the strongest correlation is observed among the theories tested (R2 = 0.81, Figure 4d). Specifically, larger OADs for the high-κ layers resulted in larger ΦB values. The changes in ΦB relative to the SiOx control samples are also consistent with the OAD model. Assuming a stoichiometry of SiO2 and a film density of 2.2 g/cm3, the interfacial SiO2 would have an OAD of approximately 1.2 × 1015 cm–2.19 All of the films with higher OADs than that of SiO2 exhibit an increase in ΦB upon introduction of the high-κ layer, and both of those with a smaller OAD exhibit decrease.

Figure 4.

Figure 4

Open in a new tab

Schottky barrier height shifts quantified from C–V data plotted against the corresponding high-κ oxide’s (a) parent metal work function, (b) parent metal Pauling electronegativity, (c) measured dielectric constant, and (d) estimated OAD. The arrow in (d) indicates the approximate OAD of SiOx, where the ΦB shift would be expected to be zero.

A notable exception to the trend of the OAD versus ΦB is that the addition of SrOx and LaOx induces similar ΦB shifts despite the significant difference in the estimated OAD between them. Atomic ratios determined by XPS measurements indicated partial conversion of the films to SrCO3 and La(OH)3, whereas the compositions used in the calculations were SrO and La2O3. Hence, the OADs estimated for these films may be less accurate than those for the other oxides. A separate plot (not shown) was made for adjusting the formulas in Table 2 based on the atomic ratios measured via XPS. While this approach resulted in a smaller OAD difference between the Sr and La species, it yielded a poorer linear fit overall (R2 = 0.65). The poor fit is attributed to the difficulty separating contributions in XPS O 1s core level spectra from metal oxides, metal hydroxides, and adventitious contamination, which could lead to erroneous oxygen/metal atomic ratios. Even within a given oxide composition, changes in ΦB are expected based on thermal processing, native oxide characteristics, and film thickness, such that a perfect correlation of OAD differences and ΦB are not expected.3,13,18 LaOx deposited on Si with the ALD process used in the present study has also been suggested to form an interfacial silicate, and as a result, there may not be a clear SiOx/high-κ interface with a single dipole.36,37

Fermi level depinning was also considered as a possible mechanism for the observed barrier height changes. Assuming a work function of 5.2 eV for Ni, the unpinned ΦB for n-Si/Ni would be ∼1.1 eV.35 According to the metal-induced gap states (MIGS) theory of Fermi level pinning, the introduction of an interfacial oxide reduces pinning by reducing the overlap between the wave function of electrons in the metal and the semiconductor surface. Such an overlap creates energy states within the semiconductor band gap.38,39 In this case, it would be expected that ΦB would increase with the introduction of any high-κ tunnel oxide as it approaches the unpinned value, which was not observed. Thus, while Fermi level depinning in addition to a dipole contribution cannot be ruled out, Fermi level depinning alone is not a likely explanation for the observed results.

Another theory suggests that tunnel oxides passivate intrinsic defect states at the interface to reduce pinning.40Dit vs surface potential data (Figure S4), calculated from representative C–V data from each MOSCAP, did not display any systematic differences between high-ΦB oxides (i.e., AlOx, HfOx) and low-ΦB oxides (LaOx, SrOx). The AlOx and SrOx Dit-surface energy curves closely overlapped through most of their ranges. These results suggest that ΦB modulation is not related to the passivation of the interfacial states.

ΦB shifts were also plotted against several other parameters, including the parent metal valence, film stoichiometry (i.e., atomic ratios), oxide fixed charge (estimated from measured MOSCAP Vfb), and conduction and valence band edge positions below the vacuum level (data not shown). The band edge positions were determined from literature electron affinity and band gap data.41−52 No correlations were observed in any of these cases. Thus, of the parameters investigated, the estimated OAD provides the most accurate prediction of the change in ΦB when high-κ oxides are added to a Si/SiOx/Ni stack. Based on the C–V results, the ΦB shift when adding a high-κ oxide to the tunnel stack is approximately equal to

graphic file with name el3c01231_m004.jpg 4

OADs also produced the strongest correlation with the I–V–T data with an R2 value of 0.76, and the linear fit

graphic file with name el3c01231_m005.jpg 5

On other semiconductors, the introduction of tunnel insulators has been shown to result in a similar or identical ΦB shift when different metal top contacts are used.39,53 Therefore, it is expected that trends similar to those observed in the present study will apply to other semiconductor/metal systems with an interfacial SiOx/high-κ oxide tunnel stack.

The data presented support the hypothesis that the effect of high-κ tunnel oxides on ΦB is dominated by the OAD-induced dipoles, though other effects on ΦB may exist. The knowledge that the OAD can be used to predict changes in ΦB will help guide the material choice in the design of future devices. For example, a high-OAD tunnel oxide such as AlOx or HfOx can be inserted to maximize ΦB at a metal/n-Si contact or minimize ΦB at a metal/p-Si contact. Conversely, a low-OAD oxide such as LaOx or SrOx will minimize ΦB at a metal/n-Si contact or maximize ΦB at a metal/p-Si contact.

Conclusions

A systematic study has been conducted on the relationship between high-κ tunnel oxide composition and ΦB at n-Si/SiOx/high-κ/Ni contacts to test different hypotheses regarding the source of ΦB modulation. The changes in ΦB were measured when adding each of seven different high-κ tunnel oxides to the interface. The results were compared to several different factors proposed to cause the interfacial dipoles that result in Vfb or ΦB changes. The data presented correlate most closely with the model of a dipole at the SiOx/high-κ interface induced by the difference in the OAD, demonstrating that this is the best available model to predict the effect of such tunnel stacks on ΦB. The relationship between the oxide OAD and the change it induces in ΦB was quantified in a linear fit to help guide the design of future devices where control over ΦB is beneficial, for example, minimizing or eliminating ΦB for transistor source-drain contacts or maximizing ΦB for charge carrier separation in photovoltaics and power electronics.2,7,8

Acknowledgments

The authors thank Dr. Ryan Thorpe for the collection of XPS data and Md. Istiaque Chowdhury for the collection of XRR data.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsaelm.3c01231.

  • Process flow diagram illustrating the steps of diode fabrication, representative capacitance–voltage (C–V) data and corresponding Mott–Schottky plots corrected for excess capacitance from diodes with each oxide insulator, saturation current density plotted against the C–V and I–V–T-derived Schottky barrier height values with linear best-fit lines, and plot of the density of interface states (Dit) vs the surface potential calculated from representative MOSCAP C–V data for each high-κ oxide (PDF)

Author Contributions

All authors contributed equally to this work.

This research was funded by NSF grant no. 1605129.

The authors declare no competing financial interest.

Supplementary Material

el3c01231_si_001.pdf (522.8KB, pdf)

References

  1. Schroder D. K.; Meier D. L. Solar Cell Contact Resistance—A Review. IEEE Trans. Electron Devices 1984, 31 (5), 637–647. 10.1109/T-ED.1984.21583. [DOI] [Google Scholar]
  2. Connelly D.; Faulkner C.; Grupp D. E. Optimizing Schottky S/D Offset for 25-Nm Dual-Gate CMOS Performance. IEEE Electron Device Lett. 2003, 24 (6), 411–413. 10.1109/LED.2003.813363. [DOI] [Google Scholar]
  3. Coss B. E.; Loh W. Y.; Wallace R. M.; Kim J.; Majhi P.; Jammy R. Near Band Edge Schottky Barrier Height Modulation Using High- κ Dielectric Dipole Tuning Mechanism. Appl. Phys. Lett. 2009, 95 (22), 222105. 10.1063/1.3263719. [DOI] [Google Scholar]
  4. Neusel L.; Bivour M.; Hermle M. Selectivity Issues of MoOx Based Hole Contacts. Energy Procedia 2017, 124, 425–434. 10.1016/j.egypro.2017.09.268. [DOI] [Google Scholar]
  5. Garland B. M.; Davis B. E.; Strandwitz N. C. Investigating the Effect of Aluminum Oxide Fixed Charge on Schottky Barrier Height in Molybdenum Oxide-Based Selective Contacts. Sol. Energy Mater. Sol. Cells 2023, 262, 112537. 10.1016/j.solmat.2023.112537. [DOI] [Google Scholar]
  6. Pulfrey D. L.; McOuat R. F. Schottky-Barrier Solar-Cell Calculations. Appl. Phys. Lett. 1974, 24 (4), 167–169. 10.1063/1.1655138. [DOI] [Google Scholar]
  7. Lam Y. W. Min m.i.s. Schottky Barrier Solar Cells—a Review. Radio Electron. Eng. 1981, 51 (9), 447–454. 10.1049/ree.1981.0068. [DOI] [Google Scholar]
  8. Sun Y.; Kang X.; Zheng Y.; Lu J.; Tian X.; Wei K.; Wu H.; Wang W.; Liu X.; Zhang G. Review of the Recent Progress on GaN-Based Vertical Power Schottky Barrier Diodes (SBDs). Electronics 2019, 8 (5), 575. 10.3390/electronics8050575. [DOI] [Google Scholar]
  9. Schroder D. K.Semiconductor Material and Device Characterization, 3rd ed.; John Wiley & Sons, Inc., 2006; Vol. 44. [Google Scholar]
  10. Yeo Y. C.; Ranade P.; King T. J.; Hu C. Effects of high-/spl kappa/gate dielectric materials on metal and silicon gate workfunctions. IEEE Electron Device Lett. 2002, 23 (6), 342–344. 10.1109/LED.2002.1004229. [DOI] [Google Scholar]
  11. Connelly D.; Faulkner C.; Clifton P. A.; Grupp D. E. Fermi-Level Depinning for Low-Barrier Schottky Source/Drain Transistors. Appl. Phys. Lett. 2006, 88, 012105. 10.1063/1.2159096. [DOI] [Google Scholar]
  12. Hu J.; Nainani A.; Sun Y.; Saraswat K. C.; Philip Wong H. S. Impact of Fixed Charge on Metal-Insulator-Semiconductor Barrier Height Reduction. Appl. Phys. Lett. 2011, 99, 252104. 10.1063/1.3669414. [DOI] [Google Scholar]
  13. Marstell R. J.; Pugliese A.; Strandwitz N. C. Absence of Evidence for Fixed Charge in Metal-Aluminum Oxide-Silicon Tunnel Diodes. Phys. Status Solidi B 2019, 256, 1800342. 10.1002/pssb.201800342. [DOI] [Google Scholar]
  14. Garland B. M.; Davis B. E.; Strandwitz N. C.. Exploiting Fixed Charge to Control Schottky Barrier Height in Si|Al2O3|MoOx- Based Tunnel Diodes. 2021 IEEE 48th Photovoltaic Specialists Conference, 2021; pp 2439–2442. 10.1109/PVSC43889.2021.9519088. [DOI] [Google Scholar]
  15. Yamamoto Y.; Kita K.; Kyuno K.; Toriumi A. Study of La-Induced Flat Band Voltage Shift in Metal/HfLaO x/SiO2/Si Capacitors. Jpn. J. Appl. Phys., Part 1 2007, 46 (11R), 7251–7255. 10.1143/JJAP.46.7251. [DOI] [Google Scholar]
  16. Iwamoto K.; Kamimuta Y.; Ogawa A.; Watanabe Y.; Migita S.; Mizubayashi W.; Morita Y.; Takahashi M.; Ota H.; Nabatame T.; Toriumi A. Experimental Evidence for the Flatband Voltage Shift of High- k Metal-Oxide-Semiconductor Devices Due to the Dipole Formation at the High- KSi O2 Interface. Appl. Phys. Lett. 2008, 92, 132907. 10.1063/1.2904650. [DOI] [Google Scholar]
  17. Kirsch P. D.; Sivasubramani P.; Huang J.; Young C. D.; Quevedo-Lopez M. A.; Wen H. C.; Alshareef H.; Choi K.; Park C. S.; Freeman K.; Hussain M. M.; Bersuker G.; Harris H. R.; Majhi P.; Choi R.; Lysaght P.; Lee B. H.; Tseng H. H.; Jammy R.; Böscke T. S.; Lichtenwalner D. J.; Jur J. S.; Kingon A. I. Dipole Model Explaining High-k/Metal Gate Field Effect Transistor Threshold Voltage Tuning. Appl. Phys. Lett. 2008, 92, 092901. 10.1063/1.2890056. [DOI] [Google Scholar]
  18. Coss B. E.; Loh W. Y.; Carlo Floresca H.; Kim M. J.; Majhi P.; Wallace R. M.; Kim J.; Jammy R. Dielectric Dipole Mitigated Schottky Barrier Height Tuning Using Atomic Layer Deposited Aluminum Oxide for Contact Resistance Reduction. Appl. Phys. Lett. 2011, 99 (10), 102108. 10.1063/1.3633117. [DOI] [Google Scholar]
  19. Kita K.; Toriumi A. Origin of Electric Dipoles Formed at High- k/SiO2interface. Appl. Phys. Lett. 2009, 94 (13), 92–95. 10.1063/1.3110968. [DOI] [Google Scholar]
  20. Fujimura N.; Ohta A.; Ikeda M.; Makihara K.; Miyazaki S. Direct Evaluation of Electrical Dipole Moment and Oxygen Density Ratio at High-k Dielectrics/SiO2 Interface by X-Ray Photoelectron Spectroscopy Analysis. Jpn. J. Appl. Phys. 2018, 57, 04FB07. 10.7567/JJAP.57.04FB07. [DOI] [Google Scholar]
  21. Lin L.; Robertson J. Atomic Mechanism of Electric Dipole Formed at High-K: SiO2 Interface. J. Appl. Phys. 2011, 109 (9), 094502. 10.1063/1.3583655. [DOI] [Google Scholar]
  22. Zhu L. Q.; Barrett N.; Jégou P.; Martin F.; Leroux C.; Martinez E.; Grampeix H.; Renault O.; Chabli A. X-Ray Photoelectron Spectroscopy and Ultraviolet Photoelectron Spectroscopy Investigation of Al-Related Dipole at the HfO2/Si Interface. J. Appl. Phys. 2009, 105, 024102. 10.1063/1.3066906. [DOI] [Google Scholar]
  23. Coss B. E.; Sivasubramani P.; Brennan B.; Majhi P.; Wallace R. M.; Kim J. Measurement of Schottky Barrier Height Tuning Using Dielectric Dipole Insertion Method at Metal-Semiconductor Interfaces by Photoelectron Spectroscopy and Electrical Characterization Techniques. J. Vac. Sci. Technol., B: Nanotechnol. Microelectron.: Mater., Process., Meas., Phenom. 2013, 31 (2), 021202. 10.1116/1.4788805. [DOI] [Google Scholar]
  24. Kern W. The Evolution of Silicon Wafer Cleaning Technology. J. Electrochem. Soc. 1990, 137 (6), 1887–1892. 10.1149/1.2086825. [DOI] [Google Scholar]
  25. Gelderman K.; Lee L.; Donne S. W. Flat-Band Potential of a Semiconductor: Using the Mott-Schottky Equation. J. Chem. Educ. 2007, 84 (4), 685–688. 10.1021/ed084p685. [DOI] [Google Scholar]
  26. Bhuiyan A. S.; Martinez A.; Esteve D. A New Richardson Plot for Non-Ideal Schottky Diodes. Thin Solid Films 1988, 161, 93–100. 10.1016/0040-6090(88)90239-8. [DOI] [Google Scholar]
  27. Terman L. M. An Investigation of Surface States at a Silicon/Silicon Oxide Interface Employing Metal-Oxide-Silicon Diodes. Solid-State Electron. 1962, 5, 285–299. 10.1016/0038-1101(62)90111-9. [DOI] [Google Scholar]
  28. Groner M. D.; Fabreguette F. H.; Elam J. W.; George S. M. Low-Temperature Al2O3 Atomic Layer Deposition. Chem. Mater. 2004, 16 (4), 639–645. 10.1021/cm0304546. [DOI] [Google Scholar]
  29. Lee B. H.; Anderson V. R.; George S. M. Growth and Properties of Hafnicone and HfO2/Hafnicone Nanolaminate and Alloy Films Using Molecular Layer Deposition Techniques. ACS Appl. Mater. Interfaces 2014, 6 (19), 16880–16887. 10.1021/am504341r. [DOI] [PubMed] [Google Scholar]
  30. Abendroth B.; Moebus T.; Rentrop S.; Strohmeyer R.; Vinnichenko M.; Weling T.; Stöcker H.; Meyer D. C. Atomic Layer Deposition of TiO2 from Tetrakis(Dimethylamino) Titanium and H2O. Thin Solid Films 2013, 545, 176–182. 10.1016/j.tsf.2013.07.076. [DOI] [Google Scholar]
  31. Vasudev P. K.; Mattes B. L.; Pietras E.; Bube R. H. Excess Capacitance and Non-Ideal Schottky Barriers on GaAs. Solid-State Electron. 1976, 19, 557–559. 10.1016/0038-1101(76)90052-6. [DOI] [Google Scholar]
  32. Tung R. T. schottky barrier heights of single crystal silicides on si(111). J. Vac. Sci. Technol., B: Microelectron. Nanometer Struct.—Process., Meas., Phenom. 1984, 2 (3), 465–470. 10.1116/1.582896. [DOI] [Google Scholar]
  33. Werner J. H.; Güttler H. H. Barrier Inhomogeneities at Schottky Contacts. J. Appl. Phys. 1991, 69 (3), 1522–1533. 10.1063/1.347243. [DOI] [Google Scholar]
  34. Singhania R. M.; Price H.; Kounga V. Y.; Davis B.; Brüner P.; Thorpe R.; Hynek D. J.; Cha J. J.; Strandwitz N. C. Surface Characterization of Ultrathin Atomic Layer Deposited Molybdenum Oxide Films Using High-Sensitivity Low-Energy Ion Scattering. J. Vac. Sci. Technol., A 2021, 39 (6), 063210. 10.1116/6.0001164. [DOI] [Google Scholar]
  35. CRC Handbook of Chemistry and Physics, 95th ed.; Haynes W. M.; Lide D. R.; Bruno T. J., Eds.; CRC Press: Boca Raton, FL, 2014. [Google Scholar]
  36. Park T. J.; Byun Y. C.; Wallace R. M.; Kim J. Impurity and Silicate Formation Dependence on O3 Pulse Time and the Growth Temperature in Atomic-Layer-Deposited La2O3 Thin Films. J. Chem. Phys. 2017, 146, 052821. 10.1063/1.4975083. [DOI] [PubMed] [Google Scholar]
  37. Davis B. E.; Strandwitz N. C.. Atomic Layer Deposited Bilayers and the Influence on Metal-Insulator-Semiconductor Schottky Barriers. 2022 IEEE 49th Photovoltaics Specialists Conference, 2022; pp 1323–1328. 10.1109/pvsc48317.2022.9938608. [DOI]
  38. Zhou Y.; Ogawa M.; Han X.; Wang K. L. Alleviation of Fermi-Level Pinning Effect on Metal/Germanium Interface by Insertion of an Ultrathin Aluminum Oxide. Appl. Phys. Lett. 2008, 93 (20), 91–94. 10.1063/1.3028343. [DOI] [Google Scholar]
  39. Nishimura T.; Kita K.; Toriumi A. A Significant Shift of Schottky Barrier Heights at Strongly Pinned Metal/Germanium Interface by Inserting an Ultra-Thin Insulating Film. Appl. Phys. Express 2008, 1, 051406. 10.1143/APEX.1.051406. [DOI] [Google Scholar]
  40. Zhou Y.; Han W.; Wang Y.; Xiu F.; Zou J.; Kawakami R. K.; Wang K. L. Investigating the Origin of Fermi Level Pinning in Ge Schottky Junctions Using Epitaxially Grown Ultrathin MgO Films. Appl. Phys. Lett. 2010, 96, 102103. 10.1063/1.3357423. [DOI] [Google Scholar]
  41. Rangan S.; Bersch E.; Bartynski R. A.; Garfunkel E.; Vescovo E. Band Offsets of a Ruthenium Gate on Ultrathin High- κ Oxide Films Phys. Rev. B: Condens. Matter Mater. Phys.on Silicon. Phys. Rev. B - Condens. Matter Mater. Phys. 2009, 79, 075106. 10.1103/PhysRevB.79.075106. [DOI] [Google Scholar]
  42. Yin X.; Battaglia C.; Lin Y.; Chen K.; Hettick M.; Zheng M.; Chen C.; Kiriya D.; Javey A. 19.2% Efficient InP Heterojunction Solar Cell with Electron-Selective TiO2 Contact. ACS Photonics 2014, 1, 1245–1250. 10.1021/ph500153c. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Tsou K. Y.; Hensley E. B. Electron Affinities of the Alkaline Earth Chalcogenides. J. Appl. Phys. 1974, 45 (1), 47–49. 10.1063/1.1663005. [DOI] [Google Scholar]
  44. Kadir R. A.; Rani R. A.; Zoolfakar A. S.; Ou J. Z.; Shafiei M.; Wlodarski W.; Kalantar-Zadeh K. Nb2O5 Schottky Based Ethanol Vapour Sensors: Effect of Metallic Catalysts. Sens. Actuators, B 2014, 202, 74–82. 10.1016/j.snb.2014.04.083. [DOI] [Google Scholar]
  45. Filho L. F.; Pansini F. N. N.; de Souza F. A. L. Size and shape effects on the stability, electronic structure, and Raman spectroscopy of (SrO)n nanoclusters. Int. J. Quantum Chem. 2021, 121 (12), e26642 10.1002/qua.26642. [DOI] [Google Scholar]
  46. Kandpal K.; Gupta N. Investigations on High-κ Dielectrics for Low Threshold Voltage and Low Leakage Zinc Oxide Thin-Film Transistor, Using Material Selection Methodologies. J. Mater. Sci.: Mater. Electron. 2016, 27 (6), 5972–5981. 10.1007/s10854-016-4519-0. [DOI] [Google Scholar]
  47. Filatova E. O.; Konashuk A. S. Interpretation of the Changing the Band Gap of Al2O3 Depending on Its Crystalline Form: Connection with Different Local Symmetries. J. Phys. Chem. C 2015, 119 (35), 20755–20761. 10.1021/acs.jpcc.5b06843. [DOI] [Google Scholar]
  48. Alimardani N.; King S. W.; French B. L.; Tan C.; Lampert B. P.; Conley J. F. Investigation of the Impact of Insulator Material on the Performance of Dissimilar Electrode Metal-Insulator-Metal Diodes. J. Appl. Phys. 2014, 116, 024508. 10.1063/1.4889798. [DOI] [Google Scholar]
  49. Kim G. M.; Lee S. M.; Michler G. H.; Roggendorf H.; Gösele U.; Knez M. Nanostructured Pure Anatase Titania Tubes Replicated from Electrospun Polymer Fiber Templates by Atomic Layer Deposition. Chem. Mater. 2008, 20 (9), 3085–3091. 10.1021/cm703398b. [DOI] [Google Scholar]
  50. Park C.; Lee C.; Lee W.; Lee J.; Kim J.; Eom D.; Oh J.; Lee S. H.; Kim H. Thermal Stability of MgO Film on Si Grown by Atomic Layer Deposition Using Mg(EtCp)2 and H2O. Ceram. Int. 2021, 47 (22), 31583–31589. 10.1016/j.ceramint.2021.08.038. [DOI] [Google Scholar]
  51. Hou X. Y.; Tan J.; Hu C. E.; Chen X. R.; Geng H. Y. Thermoelectric Properties of Strontium Oxide under Pressure: First-Principles Study. Phys. Lett. A 2021, 390, 127083. 10.1016/j.physleta.2020.127083. [DOI] [Google Scholar]
  52. Wang X.; Liu H.; Zhao L.; Fei C.; Feng X.; Chen S.; Wang Y. Structural Properties Characterized by the Film Thickness and Annealing Temperature for La2O3 Films Grown by Atomic Layer Deposition. Nanoscale Res. Lett. 2017, 12 (1), 233. 10.1186/s11671-017-2018-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Hu J.; Saraswat K. C.; Wong H. S. Metal/III-V Schottky barrier height tuning for the design of nonalloyed III-V field-effect transistor source/drain contacts. J. Appl. Phys. 2010, 107, 063712. 10.1063/1.3327434. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

el3c01231_si_001.pdf (522.8KB, pdf)

Articles from ACS Applied Electronic Materials are provided here courtesy of American Chemical Society

RESOURCES