AL2O3 / Zr doped HfO2 ferroelectric tunneling junction and its potential applications as memristors

AL

2

O

3

/ Zr doped HfO

2

ferroelectric

tunneling junction and its potential

applications as memristors

By

Haonan Wu

Senior Thesis in Electrical Engineering

University of Illinois at Urbana-Champaign

Advisor: Prof. Wenjuan Zhu

ii

Abstract

Ferroelectric tunneling junctions (FTJs), with tunable tunneling electroresistance (TER), are promising for many emerging applications including memristors and neurosynaptic computing. Here we propose a novel design of HZO layer based FTJ by adding an extra Al2O3 capping layer. The interfacial Al2O3 layer and device topology enable observable tunable tunneling electroresistance even when the thickness of HZO is above 10nm. Evidently the polarization switching in HZO is elusively dominated by atomic scale ferroelectric domains dispersed in the films. Moreover, the conductance variation of our device under voltages pulses is extracted from the measured I-V curves, and the analysis based on conductance variation validates the reliability and performance of fabricated FTJs. Our results reveal potential opportunities for HZO ferroelectrics as the hardware basis of future non-volatile memories.

Subject Keywords: Ferroelectricity, Zr-doped HfO2 (HZO), tunneling electroresistance (TER), atomic layer deposition (ALD), rapid thermal processing (RTP)).

iii

Acknowledgments

We would like to acknowledge support from the NSF under Grants ECCS 16-53241 CAR, Prof. Wenjuan Zhu and my collaborator Hojoon Ryu.

iv

Contents

1 Introduction ... 1

2 Literature Review ... 2

3 Description of Research Results ... 5

3.1 Device Fabrication ... 5

3.2 Fabrication Process and Tools involved ... 6

3.3 Measurements & Analysis ... 9

4. Conclusion ... 13

References... 14

Appendix A: Experiment Raw data and fitting models ... 15

1

1 Introduction

A memristor is a tunable resistor whose present resistance depends on the history of current that had previously flowed through the device. It is one of the most promising candidates that could emulate biological synapses [1,2]. Using a set of training pulses, the conductance of Memristors could be modulated continuously. Theoretical [3-6] and experimental [7-11] works have revealed TER switching in FTJs is attributed to ferroelectric modulation on barrier height.

TER switching in FTJs is attributed to ferroelectric modulation of Tunneling barrier height. Modulation of the barrier width is seriously limited in traditional FTJs, even though the tunneling transmittance depends on barrier height exponentially [12]. Also, restricted by complex perovskite systems, traditional FTJs suffer from limited CMOS-compatibility and faces severe scaling issues in contemporary Industry nodes. In contrast, doped HfO2 has the advantages of high coercive field and full compatibility with CMOS processes. Among various doped HfO2, Zr-doped HfO2 (HZO) is particularly attractive due to its low annealing temperature and excellent scalability. Moreover, electrical properties of those HZO capacitors from HfxZr1-xO2 material system, particularly the dependences of coercive field, remnant polarization, leakage current, are systematically analyzed.[22] Compared to traditional perovskites, HZO ferroelectric layer has the advantages of high coercive field and full compatibility with CMOS process [13,14].

2

2 Literature Review

In order to investigate the mechanism of the Ferroelectric Tunneling Junctions, we need to understand the physics behind FTJs. A systematic investigation of the electroresistance of Pt/BaTiO3/Nb:SrTiO3

metal/Ferro/semiconductor tunnel junction by engineering the Schottky barrier on Nb:SrTiO3 surface via varying BaTiO3 thickness and Nb doping concentration is demonstrated by [19].

In their work, the device with 0.1 wt% Nb concentration and a 4-unit-cell-thick BaTiO3 barrier has the best performance and its optimum ON/OFF ratio is as great as 6.0*106_{. The most valuable part is the} illustration of the fundamental physics behind FTJs as shown in figure 1.

Figure 1: Energy profiles at zero bias for the ON and the OFF states of the junctions with Nb concentrations of 1.0, 0.1 and 0.01 wt%, respectively, where the green and the orange arrows denote the direct tunneling and the thermally assisted tunneling, respectively. [19]

For heavily doped NbSTO with Nb concentration at 1.0 wt% shown in Fig. 1f, the Schottky barrier at the BTO/NbSTO interface is absent in the ON state. Therefore, the transport is governed by direct electron tunneling through BTO barrier (Fig. 1f). When the Nb concentration decreases to 0.1 wt% as shown in Fig. 1g, the Schottky barrier appears. This leads to the thermally assisted tunneling at ON state. The direct tunneling may be suppressed by Schottky barrier, while the thermally assisted tunneling, in which the

3

electrons are thermally emitted to an energy above the Fermi level and then tunnel through a thinner barrier, is still pronounced with this weak Schottky barrier.

Therefore, the FTJ with the Nb concentration of 0.1 wt% exhibits a large ON state current. As the Nb concentration further decreases, both Vbi and Wd increase in the ON state. The device with the Nb concentration at 0.01 wt% exhibits a weak Schottky barrier of 0.40 eV in height and 168 nm in width as shown in Fig. 1h. The thermally assisted tunneling through the barrier is hence suppressed.

Moreover, in the OFF state, the surface of heavily doped NbSTO, with the Nb concentration at 1.0 wt%, is switched to depletion. For the device with the Nb concentration at 1.0 wt%, such a thin barrier cannot suppress the thermally assisted tunneling completely. Therefore, the OFF-state current is efficiently shut off as the barrier gets high enough and wide enough to suppress the thermally assisted tunneling when the Nb concentration decreases to 0.1 wt%. Therefore, the optimum ON/OFF ratio is achieved at Nb

concentration of 0.1 wt%. The experimental data is also shown in Fig. 1a-c, which corroborates to theoretical analysis of FTJs.

Figure 2: (a)Dependence of the junction resistance measured at Vread =100mV after the application of 20 ns voltage pulses (Vwrite) of different

amplitudes. The different curves correspond to different consecutive measurements, with varying maximum (positive or negative) Vwrite. (b) Red- (and

blue-) framed images show states achieved by the application of positive (and negative) voltage pulses of increasing amplitude starting from the ON (and OFF) state. [20]

4

Meanwhile, [20] clearly shows the ferroelectric domain phase transitions obtained using PFM phase images. It is clear that the ferroelectric domain switching results in the change of junction resistance from 2 ∗ 105 _{Ω to 4 ∗ 10}7 _{Ω, which are called “ON” state and “OFF” state respectively.}

In figure 2a, voltage pulses and resistance are shown in the same plot, whereas in figure 2b, variation of capacitor resistance with the relative fraction of down domains extracted from the PFM phase images are combined and demonstrated. The blue and red symbols correspond to the experimental resistance value as a function of the fraction of down domains extracted from the PFM phase images; the black curve is a simulation in a parallel resistance model. The error bars are calculated from the distribution of clear and dark contrasts in the grey level histograms.

The reduction of the bulk characteristics of the HZO layers to minimize the paraelectric M phase is the key to stable ferroelectricity with a large remnant polarization. This idea is best demonstrated by our Zr doped HfO2 structure.

5

3 Description of Research Results

In the first part of the research, ferroelectric capacitors composed of Zr-doped HfO2 (HZO) ultrathin films (~10nm) are fabricated through atomic layer deposition (ALD) in order to observe ferroelectricity in high temperature annealed HZO layers. In order to obtain ferroelectricity in HZO ultrathin films, TiN capping layers are required to sandwich the HZO layer, and they also perform the role of capacitor electrodes. It is also found that experimental parameters such as different annealing temperature, contact electrode metals, and relative Zr composition in HZO strongly affect the leakage current and remnant polarization, whereas the coercive field was relatively fixed.

Finally, the FTJs based on Al2O3/HZO stacks were fabricated on highly doped p-Si substrates using atomic layer deposition (ALD) and they were systematically characterized. Those devices with tunable tunneling electroresistance (TER) are promising for many emerging applications including non-volatile memories and neurosynaptic computing [15].

3.1 Device Fabrication

According to previous experiments from our research group, TiN electrode is one of the necessary conditions to induce the ferroelectricity in HZO layers, and the other condition is high temperature annealing. I focused on electrode deposition of some potentially useful Ferroelectric capacitors during my sophomore year. Initially I was troubled by the contaminated TiN target, in which the deposited TiN film shows dark yellow all over the place.

After taking care of the electrode deposition, it was quite a challenge to choose an appropriate annealing temperature and period of the device. The HZO layer could break down easily while the electrodes melt, because the annealing temperature ranges from 600 °C to 1000 °C depending on the Zr composition. In my first trial, ferroelectric capacitors composed of Zr-doped HfO2 (HZO) ultrathin films (~10nm) are fabricated through atomic layer deposition (ALD), and two layers of 90 nm TiN are deposited as capping electrodes to sandwich the HZO layer. After that, the entire structure is annealed at 600-750 °C using rapid thermal processing (RTP) to activate inplane ferroelectric domains in order for ferroelectricity to be observed.

The next stage device contains an extra layer of Al2O3 of only several nanometers, which is now a Ferroelectric tunneling junction, and we no longer use TiN as electrode and capping layer. Pt top

6

electrode and p-doped silicon substrate directly as the bottom electrode at suitable annealing temperature could also activate the ferroelectricity. Our FTJ structure is shown in figure 3.

P doped Silicon Substrate

Ti/Au Alloy

Aluminum Oxide

Zr doped Hf Oxide (tenary)

Figure 3: FTJ Capacitor

3.2 Fabrication Process and Tools involved

In this section, all crucial equipment involved in the device fabrication are introduced and their working principles are briefly explained.

Atomic Layer Deposition (ALD)

The ALD process in figure 4 is based on the sequential release of gaseous precursor (chemical reactants) pulses to deposit materials on the substrate layer by layer (thickness is about one or two atoms). In the process, an initial precursor is introduced into the process chamber producing a monolayer on the substrate surface. Afterwards, the chamber is purged with an inert carrier gas to remove unreacted precursor and reaction by-products, and a second precursor is pulsed into the chamber reacting with the first precursor to produce a monolayer of the desired film on the substrate surface. This process is based on two fundamental mechanisms: chemisorption saturation process and sequential surface chemical reaction [16].

7

Figure 4: (a) ALD illustration. (b)Pictures of Sample in ALD chamber

ALD coatings provide excellent adhesion to the substrate layer, while maintaining low crystal strain. There are three main advantages:

• Perfect Films: The pinhole-free films maintain high uniformity even for large areas. The high

reliability and quality of films are readily achieved.

• Ultimate process control: Using ALD, we not only have full control on film thickness to atomic

precision (~0.3 nm), but also the composition of film material at each layer by changing the precursor at each cycle.

• Low Deposition Temperature: Because ALD only needs 150 ℃ for most depositions, the influence on the device quality is minimized, and the compatibility with other processes is extremely high.

Therefore, we could precisely control both the composition and thickness of the HZO layer (HfxZr1-xO2 material system) by tuning the number of cycles of precursor purge and type of precursor for each cycle. Each precursor will purge a layer of specific molecules is deposited. After the next cycle all molecules deposited will be broken and another type of element from the new cycle will replace one type of element in the previously formed layer of molecules to form the desired compound on the substrate.

#1 Cycle: HfCl4

#2 Cycle: H2O

H+ _{comes with cycle #2 and will bind with Cl and form HCl; Hf and O}

8

Figure 5: ALD Process Illustration [17]

Lesker Sputtering

The most crucial step after the atomic layer deposition is Plasma Sputtering of the TiN electrode. TiN layers have similar electrical properties when compared to regular metal electrodes such as Pt/Au. However, according to our previous results, the HZO layer only shows ferroelectricity after TiN capping and high temperature annealing.

The Plasma Sputtering system provided by Lesker is shown in figure 6a. There are two sputtering sources in the chamber, and the opened cap reveals our TiN target. Although the TiN target is cracked due to an accident, it still works and the TiN deposition is reliable. Our desired thickness is 90 nm, and so the deposition condition is 90 s at 300 W Ar Plasma. In addition, the plasma needs to be ignited at 50 W, and the output power needs to rise to 300 W in steps of 50 W. Once 300 W plasma is obtained, the cap of TiN target is open and the process begins.

Figure 6: Lesker RF Sputtering Setups. (a) plasma sputtering target with a open cap (TiN). (b) Sample covered with mask.

9

Ellipsometer

A spectroscopic ellipsometer in figure 7 is one of most affordable thin film characterization tools

available in most cleanrooms. It could characterize composition, roughness, thickness (depth), crystalline nature, doping concentration, electrical conductivity and other material properties [18].

Figure 7: (a) Ellipsometer mechanism and (b) commercialized ellipsometer from J.A. Woollam CO., Inc

The increasing popularity of the Ellipsometer poses new challenges to the technique, such as measurements on unstable liquid surfaces and microscopic imaging. One of the biggest challenges I encountered using the ellipsometer is the mechanical vibration of its supporting table. Because there are many large pieces of equipment operating in the same room that induce mechanical vibration; its measurement resolution is significantly limited.

Moreover, it is difficult to evaluate the thickness of new materials, such as the Hf0.5Zr0.5O2 we deposited, because its refraction index and other crucial physical parameters necessary for Ellipsometry are

unknown. So, we only use it to extract the thickness of the Al2O3 layer in the FTJ we made. And it is helpful for us to examine the uniformity of the ALD deposition.

3.3 Measurements & Analysis

The measurement results and analysis are mainly based on the two types of devices we fabricated: the ferroelectric capacitors and ferroelectric tunneling junctions (FTJs). First, we verified the ferroelectricity in 10nm HZO layer using P-E hysteresis curves, and we observed the beautiful butterfly curve we expected.

After the ferroelectricity is verified, we tested the remnant polarization when Zr composition is tuned. The remnant polarization in 50% Zr HZO layer, or Hf0.5Zr0.5O2, can reach up to ± 25 µC/cm2 with sharp switching at its coercive field as shown in figure 8b. Such a performance is promising compare to conventional perovskite ferroelectric material with much higher thickness.

10

Figure 8: (a) GIXRD patterns and (b) P-E hysteresis curves of Z12H6 (67% Zr), Z6H6 (50% Zr) and Z6H12 (33%) thin films.[22]

Based on the results Wui Chung obtained from ferroelectric capacitors, FTJs based on Al2O3/HZO stacks were fabricated on highly doped p-Si substrates using the same recipe which leads to strong remnant polarization.

Figure 9: I-V curves obtained for ON & OFF states at different cycles.

-3.5 -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 -20 0 20 40 60 80

Current (nA)

Bias Voltage (V)

11

Figure 9 shows the data of our first FTJ sample, and the best result is obtained in figure 10. I tested the sample using both -10 V and +10 V program pulses and the conductance of FTJ measured is continuously changing as the positive and negative pulses are applied repetitively.

From figure 10, It is easy to find that the initial Current Max is 3.034E-11A, while the final Current Max is to be 1.044E-9A as different number of pulses are applied. The number in the legend stands for the number of 10V pulses, and as we apply different number of pulses a family of diode I-V curve is obtained. In all cases, the standard model of diode applies, and so I extracted the conductance based on the diode equation shown below.

𝐼 = 𝐼𝑆(𝑒 𝑉𝐷

𝑛𝑉_{𝑇− 1) (1)}

where Is is the reverse bias saturation current, V_D is the voltage across the diode, VT is the thermal voltage, and n is the ideality factor, also known as the quality factor or sometimes emission coefficient.

Figure 10: Best Results summarizing all I-V curves.

-3 -2 -1 0 1 2 3 -2.00E-010 -1.00E-010 0.00E+000 1.00E-010 2.00E-010 3.00E-010 4.00E-010 5.00E-010 6.00E-010 7.00E-010 8.00E-010 9.00E-010 1.00E-009 1.10E-009

Curr

en

t (

A)

Bias Voltage (V)

12

Using the diode equation fitting, I extracted the conductance for each of the I-V curve. Figure 11 is the example of conductance extraction. When the positive 10V pulses are applied, the conductance is saturated at 1E-10 S after 100 pulse is applied, whereas the conductance is saturated at 1.40E-9 S after 10k negative 10V pulse is applied. The extracted Increase pulse # leads to decreases in ideality factor from near 50 to 10 in the diode equation fitting.

0 2000 4000 6000 8000 10000 12000 0.00E+000 1.00E-010 2.00E-010 3.00E-010 4.00E-010 5.00E-010 6.00E-010 Cond ucta nce # Pulses Conductance Positive +10V Pulse# Vs Conductance at V =3V

0 2000 4000 6000 8000 10000 12000 0.00E+000 2.00E-010 4.00E-010 6.00E-010 8.00E-010 1.00E-009 1.20E-009 1.40E-009

1.60E-009 Negative -10V Pulse# Vs Conductance at V =3V

Cond

ucta

nce

# Pulses

Conductance

Figure 11: -10V pulse train and +10V pulse train Vs extracted conductance.

Finally, I repeated the entire process illustrated above on our best sample, from obtaining the family of I-V curves and extracting conductance for each of them. As I repeat the measurement alternating from +10V pulses to -10V pulses, a clear loop is formed which verified that the entire process is stable and reversible, because the extracted conductance rises and falls as I applied positive and negative pulses consecutively. The TER of 16.85 is obtained by analyzing the ratio of maximum and minimum

conductance, and we finally demonstrate Ferroelectric Tunneling in this device.

Figure 12:conductance change vs. applied pulse trains.

1 10 100 1000 10000 0 10 20 30 40 50 60 70 80 90 100 110 120 3 Conductanc e (nS) # of Pulses #1 Initial-Positive Pulses #2 Negative Pulses #3 Positive Pulses #4 Negative Pulses

Conductance Change Vs applied Pulse Train

1

2 4

13

4. Conclusion

We found that an Al2O3 interfacial layer and the semiconducting substrate can enable the FTJs with high TER ratio. The novel FTJ will find its application in both energy efficient non-volatile memories and Moreover, we only used I-V measurements (DC sweep) to determine the conductance of the channel, in which the highest voltage we applied is 3.25V as shown in figure 11a, voltage above 4V in I-V

measurements will lead to junction breakdown, but the voltage pulse could be over 10V before the junction breakdown. Since the junction conductance increases exponentially as the increase of read voltage, we expect the on: off ratio (TERs) will reach to values comparable to traditional FTJs (1E6) demonstrated in the literature, because it increases exponentially with applied voltage.

14

References

[1] Chua, L. O. Memristor—the missing circuit element. IEEE Trans. Circuit Theory 18, 507–519 (1971).

[2] Strukov, D. B., Snider, G. S., Stewart, D. R. & Williams, R. S. The missing memristor found. Nature 453, 80–83 (2008).

[3] Tsymbal, E. Y. & Kohlstedt, H. Tunneling across a ferroelectric. Science 313, 181–183 (2006). [4] Kohlstedt, H., Pertsev, N. A., Contreras, J. R. & Waser, R. Theoretical current–voltage

characteristics of ferroelectric tunnel junctions. Phys. Rev. B 72, 125341 (2005).

[5] Zhuravlev, M. Y., Sabirianov, R. F., Jaswal, S. S. & Tsymbal, E. Y. Giant electroresistance in ferroelectric tunnel junctions. Phys. Rev. Lett. 94, 246802 (2005).

[6] Pantel, D. & Alexe, M. Electroresistance effects in ferroelectric tunnel barriers. Phys. Rev. B 82, 134105 (2010).

[7] Garcia, V. et al. Giant tunnel electroresistance for non-destructive readout of ferroelectric states. Nature 460, 81–84 (2009).

[8] Chanthbouala, A. et al. Solid-state memories based on ferroelectric tunnel junctions. Nature Nanotech. 7, 101–104 (2012).

[9] Gruverman, A. et al. Tunneling electroresistance effect in ferroelectric tunnel junctions at the nanoscale. Nano Lett. 9, 3539–3543 (2009).

[10] Crassous, A. et al. Giant tunnel electroresistance with PbTiO3 ferroelectric tunnel barriers. Appl. Phys. Lett. 96, 042901 (2010).

[11] Pantel, D. et al. Tunnel electroresistance in junctions with ultrathin ferroelectric Pb (Zr0.2Ti0.8) O3 barriers. Appl. Phys. Lett. 100, 232902 (2012).

[12] Griffiths, D. Introduction to Quantum Mechanics (Pearson Prentice Hall, 2005). [13] L. Chen et al., Nanoscale, vol. 10, p. 15826, 2018.

[14] J. Muller et al., Nano Letters, vol. 12, p. 4318, 2012.

[15] J. Muller, et.al. ECS Journal of Solid-State Science and Technology, vol. 4, p. N30, 2015. [16] S.M. George. Atomic Layer Deposition: An Overview. Chem. Rev. 2010, 110, 111–131.

[17] N. P. Dasgupta, X. Meng, J. W. Elam and A. B. F. Martinson, “Atomic Layer Deposition of Metal Sulfide Materials”, Acc. Chem. Res. 48, 341 (2015).

[18] R. M. A. Azzam and N. M. Bashara, Ellipsometry and Polarized Light, Elsevier Science Pub Co (1987.

[19] Chanthbouala, A. et al. A ferroelectric memristor. Nat. Mater. 11, 860–864 (2012). [20] T. Ikuno et al., Appl. Phys. Lett., vol. 99, 2011.

[21] B. Van Zeghbroeck, Principles of Electronic Devices, 2011.

[22] Lu, Y.W. & Shieh, Jay & Tsai, F.Y.. (2016). Induction of ferroelectricity in nanoscale ZrO2/HfO2 bilayer thin films on Pt/Ti/SiO2/Si substrates. Acta Materialia.

15

Appendix A: Experiment Raw data and fitting models

Figure 13: IV curves of the best performance sample.

Figure 14:IV curves of the first sample.

-4 -3 -2 -1 0 1 2 3 4 -10 -8 -6 -4 -2 0 2 4 6 8 10 12 Current ( nA) Bias Voltage (V) initial2 IV4 IV5 IV6 Equation y = a + b*x Plot A A A A Weight No Weighting

Intercept -4.56587E-8 ± 1.52724E- -8.29997E-9 ± 2.38902E-1 -1.27287E-8 ± 3.20605E-1 -1.36849E-8 ± 3.9962E-1 Slope 1.72928E-8 ± 4.6843E-1 3.33811E-9 ± 7.32755E-1 4.78527E-9 ± 9.83353E-1 5.13432E-9 ± 1.2257E-1 Residual Sum of Squares 6.93426E-22 1.69678E-23 3.05583E-23 4.74767E-23

Pearson's r 0.99963 0.99976 0.99979 0.99972 R-Square (COD) 0.99927 0.99952 0.99958 0.99943 Adj. R-Square 0.99853 0.99904 0.99916 0.99886 -4 -2 0 2 4 -20 -10 0 10 20 30 40 50 60

Current (nA)

Bias Voltage (V)

IV4

IV6

initial2

IV5

Intercept -1.03092E-7 ± 1.36786 -2.27696E-7 ± 5.6078 -1.29231E-8 ± 4.73388 -2.34102E-7 ± 8.3720 Slope 4.67219E-8 ± 4.19548 8.65566E-8 ± 1.7200 4.80913E-9 ± 1.45196E 8.85503E-8 ± 2.56786 Residual Sum of Squ 5.56252E-22 9.34915E-21 6.66226E-23 2.08378E-20 Pearson's r 0.99996 0.9998 0.99954 0.99958 R-Square (COD) 0.99992 0.99961 0.99909 0.99916 Adj. R-Square 0.99984 0.99921 0.99818 0.99832

16

Figure 15:IV curves of the best performance sample.

Figure 16:IV curves of a poor performance sample.

-4 -2 0 2 4 -30 -20 -10 0 10 20 30 40 50 60

Current (nA)

Bias Voltage (V)

initial2

IV4

IV5

IV6

Intercept -2.08179E-7 ± 5.72051E -3.76914E-8 ± 6.61274E- -2.87945E-8 ± 6.33308E- -2.82179E-8 ± 6.74979E-Slope 7.85562E-8 ± 1.75458E- 1.6124E-8 ± 2.02824E-1 1.10509E-8 ± 1.94247E- 1.06866E-8 ± 2.07028E-Residual Sum of Square 9.72874E-21 1.30002E-22 1.19239E-22 1.35447E-22 Pearson's r 0.99975 0.99992 0.99985 0.99981 R-Square (COD) 0.9995 0.99984 0.99969 0.99962 Adj. R-Square 0.999 0.99968 0.99938 0.99925 -4 -2 0 2 4 -20 0 20 40 60 80

Current (nA)

Bias Voltage (V)

initial2

IV4

IV5

IV6

Intercept -2.68559E-8 ± 7.43894E- -2.81096E-7 ± 2.49434E- -3.10673E-7 ± 7.35346E -2.91641E-7 ± 7.35048E Slope 1.01616E-8 ± 2.28165E-1 1.07865E-7 ± 7.65057E- 1.15846E-7 ± 2.25544E- 1.09886E-7 ± 2.25452E-Residual Sum of Squares 1.64516E-22 1.84968E-21 1.60757E-20 1.60627E-20 Pearson's r 0.99975 0.99997 0.99981 0.99979 R-Square (COD) 0.9995 0.99995 0.99962 0.99958 Adj. R-Square 0.99899 0.9999 0.99924 0.99916

17

Appendix B: Experimental Equipment Gallery

Figure 17: Leskers Sputter Chamber