Effect of surface roughness on electron work function of AZ31 Mg alloy and their correlation

Accepted Manuscript Effect of surface roughness on electron work function of AZ31 Mg alloy and their correlation Mingshan Xue, Yao Yao, Junfei Ou, Fajun Wang, Yao Qu, Wen Li PII:

S0925-8388(17)33410-2

DOI:

10.1016/j.jallcom.2017.10.007

Reference:

JALCOM 43404

To appear in:

Journal of Alloys and Compounds

Received Date: 22 July 2017 Revised Date:

30 September 2017

Accepted Date: 3 October 2017

Please cite this article as: M. Xue, Y. Yao, J. Ou, F. Wang, Y. Qu, W. Li, Effect of surface roughness on electron work function of AZ31 Mg alloy and their correlation, Journal of Alloys and Compounds (2017), doi: 10.1016/j.jallcom.2017.10.007. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Effect of surface roughness on electron work function of AZ31 Mg alloy and their correlation Mingshan Xue,∗ Yao Yao, Junfei Ou,* Fajun Wang, Yao Qu, Wen Li

RI PT

Key Laboratory for Microstructural Control of Metallic Materials of Jiangxi Province, School of Materials Science and Engineering, Nanchang Hangkong University, Nanchang 330063, P.R. China.

SC

ABSTRACT

Effect of surface geometrical configurations on electronic behaviors

M AN U

of various light alloys are of both fundamental and practical significance. In this study, the effect of surface roughness on the electron work function (EWF) of AZ31 Mg alloys was investigated by scanning Kelvin

TE D

probe and theoretical simulation. The experimental results indicated that the surface roughness had an important effect on the EWF, and the EWF was linearly increased with the increased roughness. It was mainly in that

EP

the change of the local electrostatic field near the rough surface confined

AC C

the movement and transfer of the electrons, resulting in the increase of the EWF. Based on the microcapacitor model, the linear relationship between the EWF and the surface roughness was reasonably explained, which was well consistent with the experimental data.

∗ Corresponding author. Tel.:+86 791 86453210; fax: +86 791 86453210. E-mail address: [email protected] (M. S. Xue); [email protected].

1

ACCEPTED MANUSCRIPT

I. INTRODUCTION Mechanical or electrochemical behaviors of material surfaces are very complicated because they involve various physical, chemical and

RI PT

mechanical factors [1,2]. For instance, the surface wear relies strongly on the material properties (such as hardness), surface configurations, strain, temperature, relative humidity, and so on [3,4]. From the point of view of

SC

fundamental study, how to understand the inherent correlation between

M AN U

the structures and properties of a material is a basic prerequisite for developing its applications [5]. In fact, for a certain parameter (such as surface roughness), its correlation with other behaviors (such as surface absorption, oxidation, corrosion, and wear) is also extremely complicated.

TE D

For example, the surface geometrical configurations as a structural characteristic of a material play an important role in determining its

EP

surface physical, chemical, electronic, and mechanical properties [6,7]. To be specific, the surface roughness changes the contact resistance

AC C

between different materials [ 8 ]; surface atomic arrangement and distribution affects the growth mechanisms of epitaxial films [9]; surface microstructures and coatings significantly change the corrosive rate, friction, and wetting behaviors [10,11]. However, the essential effects of surface geometrical configurations (such as surface roughness) on surface properties have not been completely understood until now. It is of much necessity and significance to study the inherent relationship between the 2

ACCEPTED MANUSCRIPT

surface geometrical configurations and surface parameters characterizing the surface physical and chemical properties. The electron work function (EWF), associated with the Fermi level

RI PT

(EF), is one of the most fundamental parameters of surface electronic properties of a material. The EWF of the materials has been widely used to characterize various behaviors such as adhesion, corrosion, wear, and

SC

interfacial bonding because of its high sensitivity to surface geometrical

M AN U

configurations [3,7,12,13]. Physically, the EWF is defined as the required minimum energy removing an electron from the solid surface to a point immediately outside the solid surface. In fact, a more physically meaningful definition of the EWF of the surface (ΦS) is strongly

TE D

associated with the surface dipole (ED) and EF, i.e., ΦS = ED - EF, as shown in Fig. 1(a). The surface dipole depends sensitively on the topmost surface layers exposing in air, such as surface defects, surface oxidation,

EP

and so on [14]. Thereby, the value of the EWF of a material reflects how

AC C

hard an electron transfers and transports, and the related study is of great significance for the understanding of a wide range of surface phenomena such as surface oxidation, corrosion and wear [15,16]. Some groups found that the variation of the EWF was associated with the formation of new surfaces induced by deformation [15- 17 ], but another group suggested that the variation was dependent on the roughness of deformed surface [18]. In our previous study, the decrease of the EWF of Al-Mg 3

ACCEPTED MANUSCRIPT

alloys with the increase of surface smoothness as well as the effect of surface environment on the EWF has been also reported [19]. Moreover, the further study on the field will be significant to reveal the inherent

RI PT

mechanisms responsible for the effect of surface conditions on the surface electronic behaviors.

Mg alloys, as one of the lightest structural materials with excellent

SC

physical and mechanical properties (such as high specific strength, high

M AN U

damping characteristics, and well electromagnetic shielding properties) have received considerable attention in the fields of aerospace, telecommunications, portable computers, and automobile industries, etc [20,21]. In spite of these excellent properties and wide applications, the

TE D

poor corrosive and abrasive resistance confines the use of Mg alloys. In order to improve the surface corrosion and abrasion resistance, some groups have investigated the surface modification of Mg alloys (such as

EP

the plasma electrolytic oxidation), exhibiting the better performance than

AC C

some traditional methods of surface treatments (such as anodizing) [22,23]. However, the physical mechanisms of corrosive and abrasive resistance of Mg alloys, especially their correlation with surface geometrical configurations and electronic structures are far from understanding. In the previous study [19], it was found that the EWF of Al-Mg alloys was an exceedingly sensitive parameter to the surface 4

ACCEPTED MANUSCRIPT

microstructures, and could be used to characterize the surface electronic behaviors of a material. In this paper, we focus on the effect of surface roughness on the EWF of AZ31 Mg alloys from the point of theoretical

RI PT

and experimental view. The experimental results indicated that the surface roughness had an important effect on the EWF, and the EWF was linearly increased with the increased roughness. It was mainly in that the

SC

change of the local electrostatic field near the rough surface confined the

EWF.

II. EXPERIMENT

M AN U

movement and transfer of the electrons, resulting in the increase of the

TE D

A. Surface treatment of the samples

The contents of Mg, Al, Zn, Mn in AZ31 Mg alloys (commercially available) are 95.6-95.8, 3-3.2, 0.8, 0.4 wt.%, respectively. All the

EP

specimens of AZ31 Mg alloys have the size of 15×15×3 mm3. In order to

AC C

obtain various surfaces with different roughness, they were abraded step by step using #240, #400, #600, #800, #2000 sand paper (abrasive silicon carbide, marked as samples 1, 2, 3, 4, 5). Then, these specimens were cleaned using a ultrasonic cleaner with reagent grade alcohol and acetone for 10 min, respectively. In order to reduce the surface oxidation extent, they were always immerged into the reagent grade alcohol instead of purified water with reactive oxygen, until these samples were drawn out 5

ACCEPTED MANUSCRIPT

(dried in air) for the experimental measurements. During the sample preparations and measurements, some surface phenomena (such as surface oxidation and adsorption) might not be completely avoided,

RI PT

compared with the experiments under ultrahigh vacuum conditions. However, all the samples were polished, preserved and measured under the same experimental environment, which went to the greatest extents to

SC

reduce the effect of experimental processes on the measurement results.

M AN U

B. Measurement of surface roughness and EWF

The field emission scanning electron microscope (FESEM, Nova NanoSEM 450, FEI) as well as attached energy dispersive spectroscopy (EDS) with an accelerating voltage of 20 kV were used to monitor the

TE D

surface morphology and chemical composition of these samples. The roughness profilometer (JB-6CA, HMCT) was used to measure the surface roughness of the samples, which indicated the arithmetical mean

EP

deviation (Ra) of sample outlook.

AC C

The scanning Kelvin probe (SKP, RHC020, KP Technology Ltd, Caithness, UK) was used to determine the EWF of the samples with different surface roughness. As a non-destructive surveying instrument, the principle of the SKP is exquisitely simple: to form capacitors, to allow electronic conduction, and to detect the charge transfer [19]. A three-dimensional micro-stepper positioner permitted a high-resolution sample positioning (400 nm/step). By pre-installing the peak-to-peak 6

ACCEPTED MANUSCRIPT

output voltage of the elastic peak (presenting at the output of the variable gain voltage amplifier) to be a constant, the space between the probe tip and the tested surface could be controlled within 40 nm, which ensured

RI PT

the accurate measurement of the EWF by means of the gold tip with 1 mm diameter and the shield back to the ground. Therefore, the SKP has a high sensitivity to the change in EWF (<10 meV) and has a high spatial

SC

resolution (a few tens of nanometers). The measurements of the EWF

M AN U

could be done using a point or plane scanning mode. In this work, an area of interest was scanned by the SKP and the EWF measured was the average value relative to the tip rather than the absolute value. In order to obtain the absolute EWF, the tip calibration procedure should been done

TE D

by measuring the WF difference relative to two reference samples (pure Au and Al). Based on the EWF of the Au tip is 5.1±0.1 eV and the EWF difference of Au and Al was 1.0 eV, the absolute value of the EWF is

AC C

EP

equal to the EWF of the gold tip plus the relative EWF measured.

. RESULTS AND DISCUSSION

A. Influence of surface roughness on the EWF Figure 2 shows the corresponding SEM images of AZ31 Mg alloys

after polishing using different sand paper. In order to obtain various rough surface, the samples are abraded along a certain direction. As shown in Fig. 2(a) and (b), the rougher sand paper corresponds to the 7

ACCEPTED MANUSCRIPT

rougher surface. The measured surface roughness (Ra) of Mg alloys after polishing using 240#, 400#, 600#, 800#, and 2000# sand paper is 1.220, 0.823, 0.401, 0.183, and 0.117 µm, respectively. This is to say, with the

which is in accord with the SEM results observed.

RI PT

increase of the granularity of the sand paper Ra is gradually increased,

Figure 3(a) shows the two-dimentional EWF of Mg alloys

SC

measured by SKP after polishing with 2000# sand paper (corresponding

M AN U

to Fig. 2(b)). It is observed that the distribution of the EWF in the scanning region is relatively uniform and the mean EWF measured is 4.28 eV. Figure 3(b) shows the change trend of the EWF with the increase of surface roughness. Relative to a smooth surface, the change

TE D

extent of the EWF of Mg alloys is obviously larger for the rougher surface. Compared with the effect of the surface roughness, the effect of other factors (such as SiC particles left at the surface after polishing and

EP

cleaning) on the EWF should be negligible because all the samples are

AC C

prepared under the same experimental condition. According to the least-square fitting method and the experimental data [24], the regression equation of surface roughness and EWF of Mg alloys can be expressed as: ΦS=0.3413Ra+4.2747. Namely, the measured EWF is linearly increased with the increase of surface roughness, as shown in the red line of Fig. 3(b). B. Influence of surface electrostatic field on the EWF 8

ACCEPTED MANUSCRIPT

Although the EWF is theoretically defined as the minimum energy required for extracting an electron from within the sample to a position just outside the sample, the EWF contains two parts: the chemical work

RI PT

and the electrostatic work to transport the electron through the surface [25,26]. The former is the inherent property of the material, whereas the latter is strongly associated with the surface microstructures (such as

SC

surface roughness) which readily change the local electrostatic field of

M AN U

the surface. Therefore, for a certain material, the change of the EWF is mainly affected by the distribution of the surface electrostatic potential. The SKP technique is developed based on the electrostatic drive and the low capacitance detection technique [27,28]. The SKP can be

TE D

regarded as consisting of a flat electrode (reference electrode, the Kelvin probe) suspended above and parallel to a stationary electrode (the sample measured), forming a simple parallel plate capacitor, as shown in Fig.

EP

1(b). During testing process, it can sensitively monitor the change of the

AC C

local electrostatic field between two plates, which is strongly associated with the change of surface morphology and affects the surface EWF. When the sample surface is smooth, the net charge distributes

symmetrically on the surface, generating evenly distributed symmetric field. While the sample surface is rough, more charges readily aggregate at the prominent region just as the lightning rod effect [29]. Moreover, based on the Fowler-Nordheim field emission theory [30], the 9

ACCEPTED MANUSCRIPT

electrostatic field is enhanced at sharp tips compared with valley/flat surfaces. For the rough surface, the bulgy part brings a weak repulsive force, resulting in that more charges are readily aggregated at the bulgy

RI PT

place. Therefore, the existence of the surface roughness of the sample changes the electrostatic potential distribution and electric field distribution between the Kelvin probe and the sample surface, resulting in

SC

that the electric field lines near the rough surface will be bent, as shown

M AN U

in Figure 1(b). Considering the effect of the coulomb force, the net charge will move towards the direction with smaller repulsive force. Thus, the rougher the surface is, the greater the electric potential is, the stronger the electrical field is, which is also consistent with the Fowler-Nordheim

TE D

field emission theory. Accordingly, the EWF increases because the movement of the electrons must overcome the role of the electric field. Assuming one atom has n electrons, En and En-1 are the primary

EP

energy and the final energy after one electron escapes to the region where

AC C

the electrostatic potential is Φe, the EWF can be expressed as ΦS = (En-1+Φe)-En

(1)

If the influence of the polarization after one electron escapes from the system can be ignored, so the Fermi level (EF ) and the chemical potential are approximately equal (at T=0 K, the chemical potential µ=En-En-1), the EWF of the system can be expressed as ΦS =Φe - EF 10

(2)

ACCEPTED MANUSCRIPT

For a certain material, EF is constant, and the change of the surface morphology has a tiny effect on EF. Owing to the influence of the electrostatic field, the rougher the surface is, the larger the electrostatic

RI PT

potential Φe near the rough surface is. Accordingly, the movement and transfer of the electrons are trapped and the required energy escaping from the surface increases, namely, the EWF increases with the increase

SC

of the surface roughness.

M AN U

C. Theoretical consideration of the surface roughness on the EWF From the point of the view of the SKP, the EWF is strongly dependent on the difference of the electronic energy levels between two electrodes/plates. If an external electrical contact between two electrodes

TE D

occurs, their Fermi levels become equal and the inducing movement of the electrons brings a contact potential difference (CPD). If the EWF of the Kelvin probe (ΦT) is known, then the EWF of the sample surface is

EP

related to the CPD (V) [25]:

(3)

AC C

e V=ΦT-ΦS

Where e is the charge of an electron. For a parallel plate capacitor, there is a simple correlation between the capacitance C and the contact potential difference: V=

ε ε A e , C= r 0 C D

(4)

Where εr is the relative dielectric constant, ε0 is the vacuum dielectronic constant, A is the area of the reference electrode, and D is 11

ACCEPTED MANUSCRIPT

the distance between the Kelvin probe and the sample surface (as shown in Fig. 4), then the relative EWF to the probe can be expressed as e2 D Φwf=ΦS-ΦT= ε rε 0 A

RI PT

(5)

It can be seen that εrε0A is a constant, the EWF will be mainly decided by

SC

the distance D between the Kelvin probe and the sample surface. The EWF decreases with the increase of the effective distance D.

M AN U

For the simplicity of mathematical treatment, Figure 4 simulates the rough surface with different slope. In [0, l], the slope of types 1, 2, 3 is k1, k2, k3 (k1>k2>k3), namely, the surface roughness increases with the increase of the slope k. As shown in Fig. 4, assuming the effective

TE D

distance D can be expressed as:

D = H-f(x)

(6)

EP

Where H is the constant distance between the Kelvin probe and the

AC C

sample surface. The function f(x) can be expressed as f(x )=

kx-2nkl

x∈[2nl,2nl+l]

-kx+2nkl+2kl

x

[2nl+l,2nl+2l]

n=0, 1, 2, 3…

For example, taking the interval [l, 3l] as the research scope. When x ∈ [l, 2l],

f(x)=-kx+2kl, D = kx-2kl+H

(7)

When x ∈ [2l, 3l], f (x) = kx-2kl, D = -kx+2kl+H

12

(8)

ACCEPTED MANUSCRIPT

In the interval [l, 3l], the integral for D

∫

3l

l

1 2 D dx=( kx -2klx+Hx) 2

2l l

1 2

+(- kx2+2klx+Hx)

=-kl2+2Hl

3l 2l

(9)

∫

3l

l

3l

φwf dx = ∫ − l

RI PT

Further in the interval [l, 3l], the integral for the relative EWF e2 D dx ε rε 0 A

=−

l

ε rε 0 A

SC

3l

e 2 ∫ Ddx

]

M AN U

 e 2l 2   2e 2 Hl   −   = k [  ε ε A ε ε A  r 0   r 0 

(10)

It can be seen that e2l2/εrε0A and 2e2Hl/εrε0A are constant, the greater the value of k is, the greater the EWF is. In all, the EWF increases with the

TE D

increasing surface roughness, which is well consistent with the experimental data as shown in Fig. 3(b).

EP

. CONCLUSION

AC C

The effect of surface roughness on the EWF has been investigated by means of the SKP technique and theoretical simulation. The change law of the EWF with different roughness was found, and the mechanisms of the change of the EWF at rough surface were revealed. (1) The surface roughness had an important effect on the EWF, and the EWF was linearly increased with the increased roughness according to the least-square method. 13

ACCEPTED MANUSCRIPT

(2) The change of the electrostatic field near the rough surface confined the movement and transfer of the electrons, resulting in the increase of the required energy escaping from the surface, thereby making

RI PT

the EWF increase with the increase of the surface roughness. (3) Based on the working principle of the SKP and the microcapacitor model, the linear correlation between the EWF and the surface roughness

SC

was reasonably explained, which was well consistent with the

ACKNOWLEDGMENTS

M AN U

experimental data.

The authors acknowledge with pleasure the financial support of this

TE D

work by the Natural Science Foundation of China (Grant No. 51362023 and 51662032), the Natural Science Foundation of Department of Science and Technology of Jiangxi Province (Grant No. 20152ACB21012 and

EP

20171BAB216005), the Science Foundation for Young Scientists of

AC C

Jiangxi Province, China (Grant No. 20142BCB23016), and the Natural Science Foundation of Department of Education of Jiangxi Province (Grant Nos. GJJ150722).

14

ACCEPTED MANUSCRIPT

Captions of figures: FIG. 1. Schematic view of (a) the electron work function (EWF) considering the effect of both surface dipole and material

RI PT

properties itself at the rough surface, and (b) local surface electrostatic field between the rough sample surface and the

SC

Kelvin probe surface.

M AN U

FIG. 2. SEM images of AZ31 Mg alloy surface with different surface roughness: (a) Ra = 1.22 µm after polishing with 240# sand paper; (b) Ra = 0.117 µm after polishing with 2000# sand paper.

TE D

FIG. 3. (a) Two-dimentional SKP image of the EWF of AZ31 Mg alloys with surface roughness Ra = 0.117 µm, and (b) change trend of the EWF with the increase of surface roughness. The points and

EP

lines are from the experimental data and theoretical results based

AC C

on least-square fitting method, respectively.

FIG. 4. Schematic view of: (a) the real geometric contour and three kinds of simulative geometric contour model with different slope (k1, k2,

k3) at rough surface, and (b) the parallel plate microcapacitor composed of the flat Kelvin probe and the rough alloy surface.

15

ACCEPTED MANUSCRIPT

References [1] M. Aucouturier, E. Darque-Ceretti, Proc. Mater. Sci. 9 (2015) 31. [2] G. M. Vanacore, A. W. P. Fitzpatrick, A. H. Zewail, Nano Today, 11

RI PT

(2016) 228.

[3] M. Hanief, M. F. Wani, Mater. Lett. 176 (2016) 91.

SC

[4] H. T. Cao, X. P. Dong, Z. Pan, X. W. Wu, Q. W. Huang, Y. T. Pei, Mater. Des. 100 (2016) 223.

M AN U

[5] W. Li, D. Y. Li, Acta Mater. 54 (2006) 445. [6] L. M. Liz-Marzán, Langmuir 22 (2006) 32. [7] W. Li, M. Cai, Y. Wang, S. Yu, Scripta Mater. 54 (2006) 921. [8] B. Avasarala, P. Haldar, J. Power Sources 188 (2009) 225. [9] Y. Gy. Choi, S. Y. Shin, R. Golovchak, B. Cheong, H. Jain, J. Alloy

TE D

Comp. 686 (2016) 273.

[10] B. Ge, Z. Zhang, X. Men, X. Zhu, X. Zhou, Appl. Surf. Sci. 293 (2014) 271.

EP

[11] T. Liu, L. S. Luo, D. H. Zhang, L. Wang, X. Z. Li, R. R. Chen, Y. Q.

AC C

Su, J. J. Guo, H. Z. Fu, J. Alloy Comp. 682 (2016) 663. [12] Y. J. Yu, Y. Zhao, S. Ryu, L. E. Brus, K. S. Kim, P. Kim, Nano Lett. 9 (2009) 3430. [13] M. Li, Y. Lo, S. Liao, H. Chen, H. Chang, C. Lee, C. Kuo, Appl. Surf. Sci. 357 (2015) 539.

[14] R. Smoluchowski, Phys. Rev. 60, 661 (1941). [15] D. Yamashita, A. Ishizaki, Appl. Surf. Sci. 363 (2016) 240. [16] R. Rahemi, D. Y. Li, Scrip. Mater. 99 (2015) 41. [17] V. V. Levitin, M. I. Pravda, B. A. Serpetzky, Solid State Commun. 92 16

ACCEPTED MANUSCRIPT

(1994) 973. [18] A. L. Zharin, D. A. Rigney, Tribol. Lett. 4 (1998) 205. [19] M. S. Xue, J. Xie, W. Li, C. G. Yang, Y. L. Ai, F. J. Wang, J. F. Ou, J. P. Yao, Physica B 406 (2011) 4240. [20] C. J. Silva, A. Kula, R. K. Mishra, M. Niewczas, J. Alloy Comp. 687

RI PT

(2016) 548.

[21] L. Tang, Y. Zhao, R. K. Islamgaliev, C. Y. A. Tsao, R. Z. Valiev, E. J.

SC

Lavernia, Y. T. Zhu, Mater. Sci. Eng. A 670 (2016) 280.

[22] L. Ye, J. Hu, C. Tang, X. Zhang, Y. Deng, Z. Liu, Z. Zhou, Mater.

M AN U

Charact. 79 (2013) 1.

[23] M. Nouri, Z. Liu, D. Y. Li, X. Yan, N. Tahreen, D. Chen, Thin Solid Film 615 (2016) 29.

[24] K. Usevich, I. Markovsky, Linear Algebra Appl. 502 (2016) 243. [25] W. Li, D. Y. Li, J. Chem. Phys. 122 (2005) 064708.

TE D

[26] M. S. Xue, H. N. Wu, J. F. Ou, F. J. Wang, X. B. Li, W. Li, Z. H. Jiang, J. Appl. Phys. 111 (2012) 123714.

EP

[27] M. S. Xue, J. Xie, W. Li, F. J. Wang, J. F. Ou, C. G. Yang, C. Q. Li, Z. C. Zhong, Z. H. Jiang, J. Phys. Chem. Solids 73 (2012) 781.

AC C

[28] M. S. Xue, S. B. Peng, F. J. Wang, J. F. Ou, C. Q. Li, W. Li, J. Alloy Comp. 692 (2017) 903.

[29] M. S. Xue, W. F. Wang, F. J. Wang, J. F. Ou, W. Li, Appl. Phys. Lett. 102 (2013) 243110. [30] R. H. Fowler, L. Nordheim, Proc. R. Soc. London, Ser. A 119 (1928) 173.

17

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 1 Xue et al

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 2 Xue et al

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 3 Xue et al

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 4 Xue et al

ACCEPTED MANUSCRIPT

Highlights ► The effect of surface roughness on the EWF of AZ31 Mg alloys was

RI PT

investigated. ► The change of the local electrostatic field at rough surface resulted in the increase of the EWF.

SC

► The EWF was linearly increased with the increased roughness.

AC C

EP

TE D

EWF and surface roughness.

M AN U

►A microcapacitor model was used to explain the relationship between the