The synthesis of TiO2 thin films by the chemical spray pyrolysis method at different titanium isopropoxide (TTIP) to acetylacetone (AcacH) ratios has been shown to lead to the highest photodegradation at 1 (TTIP):8 (AcacH). These films hold promise in the field of indoor pollution treatment. Carbon incorporation into the surface and into the TiO2 lattice could be responsible for the observed performance, but the mechanism is still to be elucidated. Here, we report the correlation of contact potential difference (CPD) contrast maps as produced using Kelvin Probe Force Microscopy, and the observed functionality dependence on the TTIP to AcacH ratio. Since the CPD contrast locally provides information about the sample's Fermi level, this correlation provides a means to interpret enhanced photocatalytic activity in terms of the presence of acceptors that make possible a faster transfer of charge carriers to the surface.

The synthesis and optimization of TiO2 thin films are rapidly advancing in fields such as environmental engineering, and has important and promising applications in the field of pollution treatment.1–3 These applications rely on the principles of photocatalysis4,5 and photoreactivity.6 Applications could result in the reduction of outdoor/indoor pollution7 without compromising the optimization of thermal comfort and energy consumption.8 An optimum synthesis of the films is critical for large scale implementation9 since the methods of production must be low cost, efficient, and robust.10 Several methods for synthesis are available, i.e., chemical vapor deposition,11 sputtering, and others.12 The chemical spray pyrolysis method13 holds promise in the field of self-cleaning surfaces, but choosing a precursor solution is essential and determines the functionality of the system/film. Spiridonova et al. employed1 titanium isopropoxide (TTIP) as a titanium precursor and acetylacetone (AcacH) as a stabilizing agent for the fabrication of TiO2 films by ultrasonic spray pyrolysis (see the supplementary material for details on film preparation). For the TiO2 films obtained from different TTIP to AcacH molar ratio, the photodegradation of stearic acid (SA) was monitored. It was shown that the reaction rate constant (k) increases by approximately an order of magnitude with the increasing ratio. This was quantified by calculating the reaction rate constant of the SA via the photodegradation test, while varying the ratio of TiO2 film synthesis from 1:1 (TTIP) to 1:8 (AcacH), where an optimum was found. The authors provided an interpretation for such improvement based on (1) a variation of morphology monitored via topographic AFM measurements, (2) wettability properties monitored by means of contact angle measurements, and (3) XPS measurements that showed a possible incorporation of carbon species on the surface of the TiO2 film. In another recent study,14 it was shown that increasing the amount of AcacH in the precursor solution promotes a transition from preferential fast electron to preferential fast hole transfer toward the TiO2 surface, correlating with a strong increase in the photocatalytic decomposition rate of organic pollutants. Based on these studies, the possible mechanism for enhanced photocatalytic activity was related to passivation of electron traps at the surface, induced by incorporation of carbon impurities in TiO2 films. Incorporating carbon impurities can induce “defect states” and related modifications to the band structure of the films, thus affecting the Fermi level locally.15 For this reason, Kelvin Probe Force Microscopy (KPFM) can potentially provide information regarding the changing/pinning of the Fermi level for the TiO2 system investigated by Spiridonova et al.

Here, we characterize the surface of the TiO2 films at different ratios with the AFM by means of KPFM.16 The KPFM technique can be employed17 to produce maps18 of the contact potential difference (CPD) between the AFM tip and a sample's surface. In KPFM, the tip is grounded and the CPD is defined as CPD = (ϕtipϕS)/e, where ϕS is the work function of the surface, ϕtip is the work function of the tip, and e is the electron charge. The CPD can be identified with the necessary potential energy variations to the electron structure of the tip-sample system to produce a common Fermi level EF.15 The Fermi level is the work required to add an electron to a material,19 while the work function is the minimum energy to remove an electron from it. This relationship is commonly written20 in terms of the “vacuum” energy EVAC as a reference as φ = EVACEF, where the expression shows that the work function is simply the difference between the Fermi level and the vacuum level. While the Fermi alignment affects both tip and sample, only the Fermi level of the surface varies, i.e., the tip always remains the same during the experiment. Therefore, KPFM contrast maps, through the work function difference, provide information regarding the relative Fermi level/work function from sample to sample and from area to area on a sample's surface. Since the energy required to remove/add an electron can be provided by photon adsorption, these maps can be employed to interpret variations in photodegradation/photoreactivity. Thus, variations in CPD can be employed to discuss changes in the band structure. If photodegradation is related to Fermi level variations from sample to sample, variations in CPD should also be observed in the maps as a function of the TTIP to AcacH ratio, i.e., some form of correlation between CPD and photodegradation should be possible.

The synthesis of TiO2 films, particularly in the case of the application under consideration, requires that contaminants, or pollutants, deposited on the TiO2 surface degrade. The general principle in photodegradation is that the interaction of light with matter degrades pollutants21 where reactive oxygen species (ROS)22,23 have an important mediating role24 in degradation. In the case of the TiO2 films, and in real applications, pollutants are spontaneously deposited onto the TiO2 surface from the environment. Here, emphasis is put in ROS, a collective term, rather than in free radicals alone, i.e., any species that contains one or more unpaired electrons,22 because a variety of radical and nonradical oxidizing species might be produced by photoreaction on the TiO2 surface. In short, contaminants, from here forth more restrictively referred to as volatile organic compounds (VOCs), adsorb onto TiO2 films and these degrade by the action of a set of ROS. An investigation of the presence, or absence thereof, of specific VOCs and ROS, as well as the possibility to quantify their concentration, would be a study worth carrying out, but the authors purposefully leave this question open and out of the current investigation. The aim is to focus on what can be learnt from the KPFM CPD maps alone for the different ratios and their relation to photodegradation functionality. Still, for the sake of discussion, a hypothetical and generic mechanism is illustrated in Fig. 1 and detailed with the help of Eqs. (1) and (2) in the following paragraphs. This cartoon is meant to show the role of light to activate the complex reactions in the process of photodegradation. The degradation process can be divided into three steps.23 

FIG. 1.

Three step illustration of a possible mechanism for photodegradation, where ROS act as the degrading species. (a) Water molecules in the proximity of the surface, or weakly adhered (ad) to the surface, react with electrons emitted from the TiO2 surface which have been excited by light. (b) ROS are produced and interact with VOCs. (c) The degradation of VOCs results in gases.

FIG. 1.

Three step illustration of a possible mechanism for photodegradation, where ROS act as the degrading species. (a) Water molecules in the proximity of the surface, or weakly adhered (ad) to the surface, react with electrons emitted from the TiO2 surface which have been excited by light. (b) ROS are produced and interact with VOCs. (c) The degradation of VOCs results in gases.

Close modal

In the first step, H2O in the proximity of the surface or weakly adsorbed H2O molecules and TiO2 interact with photons of energy hf to produce ROS [Eq. (1) and Figs. 1(a) and 1(b)]. In Eqs. (1) and (2), ad means adsorbed. It is this light–matter interaction that is critical when it comes to degradation since it activates the whole process. While the mechanisms are complex, it is variations in the TTIP to AcacH ratio that should control this light–matter interaction and should be visible in terms of CPD values. No attempt is made here to fully elucidate this process chemically, and the double arrows in the reaction equations emphasize this point and point to the possible presence of intermediates. The reaction kinetics of such systems can be extremely complex to quantify and monitor, and many reactive species are possible.23 A typical example of reactive species for this system with a lifespan in the range of femtoseconds is hydroxyl ̇OH. (Chen et al. provide a rather comprehensive list of main reactions and reactive species.) In terms of photodegradation, the relevance is to be found in the final products, i.e., ROS, rather than in any intermediates while the production of these ROS is dependent on band structure and surface chemistry.

In the second step, the highly reactive agents (ROS) induce a set of reactions that lead to the degradation of VOCs ultimately resulting in the production of gases, such as CO2 [Eq. (2) and Fig. 1(c)]. This stage is relevant both from a point of view of applied surface science and theoretically, but it is also outside the scope of this work to elucidate it,

(1)
(2)

Standard KPFM was performed using the lift mode whereby a topography image was acquired first and then the cantilever was lifted to track the topography and cancel the voltage resulting from the work function difference. In the experiments, we used an MFP-3D AFM from Asylum Research. The probe was a HQ:NSC18/PT (Platinum coated) with resonant frequency ∼75 kHz and spring constant ∼2 N/m. The AFM was operated in the attractive regime in the amplitude modulation (AM) AFM mode. The lift height was set to 10 nm. The tip voltage was set to 1 V, and the electrical pass drive voltage was 3 V. Topography and corresponding CPD maps are shown in Fig. 2 for the ratios (a) 1:1, (b) 1:5, and (c) 1:8 and 1:10. Data for the 1:20 ratio were also collected, but the maps are not shown. Visual inspection alone shows that topography is not correlated with the CPD maps. Spiridonova et al. carried out photodegradation experiments by varying the TTIP to AcacH ratio and exposing the samples to visible and UVA light (see the supplementary material for details on the lamp specifications). We also conducted the experiments by exposing the samples to visible light and UVA light (not all data shown). The samples were exposed for 20 min to UVA light before taking the data. An example of topography and CPD maps under (a) and (b) visible and (c) and (d) UVA light is shown in Fig. 3 for the (a) and (c) 1:1 and (b) and (d) 1:8 ratios.

FIG. 2.

Topography (left) and CPD (right) maps for the (a) 1:1, (b) 1:5, (c) 1:8, and (d) 1:10 ratios.

FIG. 2.

Topography (left) and CPD (right) maps for the (a) 1:1, (b) 1:5, (c) 1:8, and (d) 1:10 ratios.

Close modal
FIG. 3.

CPD maps as the samples are exposed to visible light (a) and (b) and UVA (c) and (d) for the 1:1 and 1:8 ratios.

FIG. 3.

CPD maps as the samples are exposed to visible light (a) and (b) and UVA (c) and (d) for the 1:1 and 1:8 ratios.

Close modal

The full data set for the experiments that we carried out is shown in Fig. 4 as histograms for the (a) UVA and (b) visible light experiments. Spiridonova et al. further reported higher photodegradation functionality when the samples were exposed to UVA, as opposed to exposure to visible light only, by monitoring and quantifying the photodegradation rate constants k (min−1). Their data are reproduced in Fig. 5 (red markers) against the results presented in Fig. 4 in terms of CPD (blue markers). The data have been normalized for convenience, and all values used to produce Fig. 5 are shown in Tables I–IV. Since the rate constant k increases with decreasing CPD, the 1-normalized CPD values are plotted against normalized k in the figure. To normalize Fig. 5(a), the values are: k (minimum) = 0.0016 min−1, k (range) = 0.0307 min−1, VCPD (minimum) = 249 mV, and VCPD (range) = 764 mV. To normalize Fig. 5(b), the values are: k (minimum)= 0.004 min−1, k (range) = 0.239 min−1, VCPD (minimum) = 213 mV, and VCPD (range) = 710 mV. These values can be directly obtained from the respective tables (Tables I–IV). The mean values and standard deviations are shown in the figures. The key points to interpret Figs. 4 and 5 follow.

FIG. 4.

Histograms showing the CPD data obtained in the KPFM experiments for (a) UVA exposed samples and (b) samples exposed to visible light. The data correspond to the 512 × 512 pixels (262 144 points) obtained from contrast maps, i.e., those in Figs. 2 and 3.

FIG. 4.

Histograms showing the CPD data obtained in the KPFM experiments for (a) UVA exposed samples and (b) samples exposed to visible light. The data correspond to the 512 × 512 pixels (262 144 points) obtained from contrast maps, i.e., those in Figs. 2 and 3.

Close modal
FIG. 5.

Normalized CPD (blue markers) and k (red markers) values for the case where (a) the samples are exposed to (a) visible and (b) UVA light. Since k typically increases with CPD, the normalization for the CPD data is carried out as 1-CPD. The figures can be reproduced from the data provided in Tables I–IV.

FIG. 5.

Normalized CPD (blue markers) and k (red markers) values for the case where (a) the samples are exposed to (a) visible and (b) UVA light. Since k typically increases with CPD, the normalization for the CPD data is carried out as 1-CPD. The figures can be reproduced from the data provided in Tables I–IV.

Close modal
TABLE I.

CPD measurements for the case when the samples are exposed to visible light.

TTIP:AcacHCPD (V)RangeSDMin. valueMax. value
1:1 0.691 0.120 0.020 0.644 0.764 
1:5 0.464 0.068 0.008 0.424 0.492 
1:8 0.282 0.063 0.009 0.249 0.311 
1:10 0.445 0.066 0.008 0.405 0.471 
1:20 0.348 0.047 0.005 0.320 0.367 
TTIP:AcacHCPD (V)RangeSDMin. valueMax. value
1:1 0.691 0.120 0.020 0.644 0.764 
1:5 0.464 0.068 0.008 0.424 0.492 
1:8 0.282 0.063 0.009 0.249 0.311 
1:10 0.445 0.066 0.008 0.405 0.471 
1:20 0.348 0.047 0.005 0.320 0.367 
TABLE II.

CPD measurements for the case when the samples are exposed to UVA.

TTIP:AcacHCPD (V)RangeSDMin. valueMax. value
1:1 0.665 0.085 0.013 0.625 0.710 
1:5 0.443 0.061 0.008 0.409 0.471 
1:8 0.246 0.061 0.009 0.213 0.274 
1:10 0.415 0.087 0.008 0.358 0.446 
1:20 0.328 0.064 0.006 0.283 0.347 
TTIP:AcacHCPD (V)RangeSDMin. valueMax. value
1:1 0.665 0.085 0.013 0.625 0.710 
1:5 0.443 0.061 0.008 0.409 0.471 
1:8 0.246 0.061 0.009 0.213 0.274 
1:10 0.415 0.087 0.008 0.358 0.446 
1:20 0.328 0.064 0.006 0.283 0.347 
TABLE III.

Photodegradation rate constants under visible light.

TTIP:AcacHk (min−1)SD
1:1 0.001 63 0.000 17 
1:3 0.008 05 0.001 46 
1:4 0.008 19 0.001 59 
1:5 0.026 67 0.003 81 
1:8 0.032 33 0.005 90 
1:10 0.030 52 0.010 53 
1:20 0.027 11 0.002 17 
TTIP:AcacHk (min−1)SD
1:1 0.001 63 0.000 17 
1:3 0.008 05 0.001 46 
1:4 0.008 19 0.001 59 
1:5 0.026 67 0.003 81 
1:8 0.032 33 0.005 90 
1:10 0.030 52 0.010 53 
1:20 0.027 11 0.002 17 
TABLE IV.

Photodegradation rate constants under UVA light.

TTIP:AcacHk (min−1)SD
1:1 0.003 98 ⋯ 
1:3 0.012 02 0.002 83 
1:4 0.018 76 0.000 87 
1:5 0.066 22 0.012 50 
1:6 0.094 71 0.000 06 
1:7 0.108 53 0.005 64 
1:8 0.242 73 0.042 48 
1:10 0.186 22 0.051 03 
1:20 0.237 43 0.045 00 
TTIP:AcacHk (min−1)SD
1:1 0.003 98 ⋯ 
1:3 0.012 02 0.002 83 
1:4 0.018 76 0.000 87 
1:5 0.066 22 0.012 50 
1:6 0.094 71 0.000 06 
1:7 0.108 53 0.005 64 
1:8 0.242 73 0.042 48 
1:10 0.186 22 0.051 03 
1:20 0.237 43 0.045 00 
  1. First, there is a shift in CPD in all cases toward lower CPD when the samples are exposed to UVA (compare the data in Table I with that in Table II where the data are presented for all samples in terms of the mean, standard deviation, range, minima, and maxima values);

  2. Second, most importantly, the minimum values of CPD are always found for the ratio 1:8. The implication is that KPFM maps can be employed to identify photodegradation efficiency directly from CPD values.

  3. Third, the maximum values of CPD are found for the ratio 1:1. It is known25 that the ratio 1:1 is the least efficient, i.e., minimum k values. This was later confirmed by Spiridonova et al. that found the highest values at 1:8.

  4. Fourth, the shift in CPD between exposing the samples to visible or UVA light is of approximately 20–40 mV for all ratios (Fig. 4). This shift is significant since the resolution of standard KPFM measurements is typically lower than 1 mV.17,26

The absolute value for the work function of the tip, ϕtip, can be found by acquiring KPFM data from a calibration sample, such as HOPG (ϕHOPG ≈ 4.6 eV).27,28 This calibration results in φtip ≈ 5.1 eV. With this measurement, the absolute values for the different ratios can be obtained, i.e., ϕS ≈ 4.4 eV (ratio 1:1), φS ≈ 4.6 eV (ratio 1:5), ϕS ≈ 4.8 eV (ratio 1:8), ϕS ≈ 4.7 eV (ratio 1:10), and ϕS ≈ 4.8 eV (ratio 1:20). Lower CPD values correspond to higher values of work function for the samples. Thus, lower work functions for the TiO2 films correlate with higher photodegradation efficiencies. This means that lower Fermi levels are obtained when the highest functionality is seen. It is known29 that semiconductor p-doping, i.e., the production of holes or acceptors near the valence band, leads to lower Fermi levels while n-doping, i.e., the production of donors near the conduction band, leads to higher Fermi levels. A hypothesis can, thus, be brought forth whereby the lower Fermi levels are the result of hole creation, i.e., analogue to p-type doping, near the surface of the samples with highest photodegradation performance. Thus, the CPD signals obtained in our experiments might indicate the presence of faster transfer of charge carriers to the surface due to the presence of acceptors for the highest ratios. This interpretation is in good agreement with the trend in Surface PhotoVoltage Spectroscopy (SPV) signal discussed by Dittrich et al.14 Fast transfer to the surface is beneficial for enhanced photocatalytic activity.

In summary, the above four points, together with the significant agreement in trends shown between k and CPD (Fig. 5) as a function of variations of the TTIP to AcacH molar ratio, show that the CPD and, therefore, the Fermi level of the samples control the photodegradation of the TiO2 films. Clearly, while there is correlation between the CPD values and the k values, the mechanisms involved in the overall process of photodegradation are complex and remain to be elucidated. It is possible that the carbon incorporation hypothesis leads to the lower Fermi levels observed whereby photodegradation is highest, i.e., highest TTIP to AcacH ratios. In analogy to p-type doping, lower Fermi levels hint at the presence of acceptors, i.e., holes, near the surface. The presence of these acceptors makes possible a faster transfer of charge carriers to the surface. Finally, since obtaining CPD data are relatively straightforward in KPFM, these results further provide the means to rapidly characterize the surfaces for photodegradation.

See the supplementary material for details on the synthesis of TiO2 thin films and the specifications of UVA and visible light lamps used to monitor stearic acid degradation.

This study was funded by the Estonian Research Council project PRG627 “Antimony chalcogenide thin films for next-generation semi-transparent solar cells applicable in electricity producing windows,” the Estonian Centre of Excellence project TK141 (No. TAR16016EK) “Advanced materials and high-technology devices for energy recuperation systems,” and the European Union's Horizon2020 programme under the ERA Chair project 5GSOLAR Grant Agreement No. 952509. This work was supported by ERDF project Centre of Technologies and Investigations of Nanomaterials (NAMUR+, Project No. 2014-2020.4.01.16-0123).

The authors have no conflicts to disclose.

Tuza A Olukan: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Visualization (equal); Writing – original draft (equal); Writing – review and editing (equal). Jekaterina Sydorenko: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Validation (equal); Visualization (equal); Writing – review and editing (equal). Atanas Katerski: Data curation (equal); Methodology (equal); Resources (equal); Validation (equal). Mariam Al Mahri: Conceptualization (equal); Investigation (equal); Visualization (equal). Chia-Yun Lai: Conceptualization (equal); Data curation (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Visualization (equal). Abdulrahman Al-Hagri: Conceptualization (equal); Investigation (equal); Methodology (equal); Validation (equal); Writing – original draft (equal); Writing – review and editing (equal). Sergio Santos: Conceptualization (equal); Formal analysis (equal); Investigation (equal); Methodology (equal); Project administration (equal); Resources (equal); Software (equal); Supervision (equal); Validation (equal); Visualization (equal); Writing – original draft (equal); Writing – review and editing (equal). Matteo Chiesa: Conceptualization (equal); Funding acquisition (lead); Investigation (supporting); Methodology (equal); Project administration (lead); Resources (lead); Supervision (lead); Writing – original draft (equal); Writing – review and editing (equal).

The data that support the findings of this study are available within the article.

1.
J.
Spiridonova
,
A.
Katerski
,
M.
Danilson
,
M.
Krichevskaya
,
M.
Krunks
, and
I. O.
Acik
,
Molecules
24
,
4326
(
2019
).
2.
S.
Banerjee
,
D. D.
Dionysiou
, and
S. C.
Pillai
,
Appl. Catal. B
176–177
,
396
428
(
2015
).
3.
Y.
Yang
,
Q.
Lai
,
S.
Mahmud
,
J.
Lu
,
G.
Zhang
,
Z.
Huang
,
Q.
Wu
,
Q.
Zeng
,
Y.
Huang
,
H.
Lei
, and
Z.
Xiong
,
J. Membr. Sci.
645
,
120204
(
2022
).
4.
J.
Schneider
,
M.
Matsuoka
,
M.
Takeuchi
,
J.
Zhang
,
Y.
Horiuchi
,
M.
Anpo
, and
D. W.
Bahnemann
,
Chem. Rev.
114
,
9919
9986
(
2014
).
5.
M. J.
Torralvo
,
J.
Sanz
,
I.
Sobrados
,
J.
Soria
,
C.
Garlisi
,
G.
Palmisano
,
S.
Çetinkaya
,
S.
Yurdakal
, and
V.
Augugliaro
,
Appl. Catal. B
221
,
140
151
(
2018
).
6.
C.
Garlisi
,
G.
Scandura
,
G.
Palmisano
,
M.
Chiesa
, and
C. Y.
Lai
,
Langmuir
32
,
11813
11818
(
2016
).
7.
A.
Mavrikos
,
D.
Papoulis
,
N.
Todorova
,
I.
Papailias
,
C.
Trapalis
,
D.
Panagiotaras
,
D. A.
Chalkias
,
E.
Stathatos
,
E.
Gianni
,
K.
Somalakidi
,
D.
Sygkridou
, and
S.
Komarneni
,
J. Photochem. Photobiol. A
423
,
113568
(
2022
).
8.
P.
Chanklom
,
T.
Kreetachat
,
R.
Chotigawin
, and
K.
Suwannahong
,
ACS Omega
6
,
10629
10636
(
2021
).
9.
K. U.
Ambikakumari Sanalkumar
and
E.-H.
Yang
,
Cem. Concr. Compos.
115
,
103847
(
2021
).
10.
T.
Adachi
,
S. S.
Latthe
,
S. W.
Gosavi
,
N.
Roy
,
N.
Suzuki
,
H.
Ikari
,
K.
Kato
,
K.
Katsumata
,
K.
Nakata
,
M.
Furudate
,
T.
Inoue
,
T.
Kondo
,
M.
Yuasa
,
A.
Fujishima
, and
C.
Terashima
,
Appl. Surf. Sci.
458
,
917
923
(
2018
).
11.
R.
Mauchauffé
,
S.
Kang
,
J.
Kim
,
J.-H.
Kim
, and
S. Y.
Moon
,
Curr. Appl. Phys.
19
,
1296
1304
(
2019
).
12.
B.
Wei
,
J. L.
Xue
,
H. T.
Cao
,
H. G.
Li
,
F.
Wen
, and
Y. T.
Pei
,
Surf. Rev. Lett.
26
,
1950036
(
2019
).
13.
A.
Juma
,
I.
Oja Acik
,
A. T.
Oluwabi
,
A.
Mere
,
V.
Mikli
,
M.
Danilson
, and
M.
Krunks
,
Appl. Surf. Sci.
387
,
539
545
(
2016
).
14.
T.
Dittrich
,
J.
Sydorenko
,
N.
Spalatu
,
N. H.
Nickel
,
A.
Mere
,
M.
Krunks
, and
I. O.
Acik
,
ACS Appl. Mater. Interfaces
(
2022
) (unpublished).
15.
P. A.
Markeev
,
E.
Najafidehaghani
,
Z.
Gan
,
K.
Sotthewes
,
A.
George
,
A.
Turchanin
, and
M. P.
de Jong
,
J. Phys. Chem. C
125
,
13551
13559
(
2021
).
16.
Y.
Adachi
,
Y.
Sugawara
, and
Y. J.
Li
,
Nano Res.
15
,
1909
1915
(
2022
).
17.
W.
Melitz
,
J.
Shen
,
A. C.
Kummel
, and
S.
Lee
,
Surf. Sci. Rep.
66
,
1
27
(
2011
).
18.
C.
Maragliano
,
S.
Lilliu
,
M. S.
Dahlem
,
M.
Chiesa
,
T.
Souier
, and
M.
Stefancich
,
Sci. Rep.
4
,
4203
(
2015
).
19.
N. W.
Ashcroft
and
D. N.
Mermin
,
Solid State Physics
(
Cengage Learning
,
1976
).
20.
A.
Kahn
,
Mater. Horiz.
3
,
7
10
(
2016
).
21.
H. D.
Burrows
,
M.
Canle L
,
J. A.
Santaballa
, and
S.
Steenken
,
J. Photochem. Photobiol. B
67
,
71
108
(
2002
).
22.
H.
Bayr
,
Crit. Care Med.
33
,
S498
–S501 (
2005
).
23.
B.
Chen
,
C.
Zhu
,
J.
Fei
,
Y.
Jiang
,
C.
Yin
,
W.
Su
,
X.
He
,
Y.
Li
,
Q.
Chen
,
Q.
Ren
, and
Y.
Chen
,
J. Hazard. Mater.
363
,
55
63
(
2019
).
24.
C.
Challier
,
S.
Laurella
,
P.
Allegretti
,
C.
Sabini
,
L.
Sabini
,
N. A.
García
,
A.
Biasutti
, and
S.
Criado
,
Photochem. Photobiol.
94
,
1151
1158
(
2018
).
25.
A. O.
Juma
,
I. O.
Acik
,
V.
Mikli
,
A.
Mere
, and
M.
Krunks
,
Thin Solid Films
594
,
287
292
(
2015
).
26.
U.
Zerweck
,
C.
Loppacher
,
T.
Otto
,
S.
Grafström
, and
L. M.
Eng
,
Phys. Rev. B
71
,
125424
(
2005
).
27.
W.
Melitz
,
J.
Shen
,
S.
Lee
,
J. S.
Lee
,
A. C.
Kummel
,
R.
Droopad
, and
E. T.
Yu
,
J. Appl. Phys.
108
,
023711
(
2010
).
28.
Y.
Lv
,
J.
Cui
,
Z. M.
Jiang
, and
X.
Yang
,
Nanoscale Res. Lett.
7
,
659
659
(
2012
).
29.
S. M.
Sze
and
K. N.
Kwok
,
Physics of Semiconductor Devices
(
Wiley Interscience
,
2007
).

Supplementary Material