Department of Science and Technology
Institutionen för teknik och naturvetenskap
Printed Schottky Diodes based
upon Zinc Oxide Materials
Emma Persson
Printed Schottky Diodes based
upon Zinc Oxide Materials
Examensarbete utfört i Teknisk fysik
vid Tekniska högskolan vid
Linköpings universitet
Emma Persson
Handledare Negar Abdollahi Sani
Examinator Isak Engquist
Detta dokument hålls tillgängligt på Internet – eller dess framtida ersättare –
under en längre tid från publiceringsdatum under förutsättning att inga
extra-ordinära omständigheter uppstår.
Tillgång till dokumentet innebär tillstånd för var och en att läsa, ladda ner,
skriva ut enstaka kopior för enskilt bruk och att använda det oförändrat för
ickekommersiell forskning och för undervisning. Överföring av upphovsrätten
vid en senare tidpunkt kan inte upphäva detta tillstånd. All annan användning av
dokumentet kräver upphovsmannens medgivande. För att garantera äktheten,
säkerheten och tillgängligheten finns det lösningar av teknisk och administrativ
art.
Upphovsmannens ideella rätt innefattar rätt att bli nämnd som upphovsman i
den omfattning som god sed kräver vid användning av dokumentet på ovan
beskrivna sätt samt skydd mot att dokumentet ändras eller presenteras i sådan
form eller i sådant sammanhang som är kränkande för upphovsmannens litterära
eller konstnärliga anseende eller egenart.
För ytterligare information om Linköping University Electronic Press se
förlagets hemsida
http://www.ep.liu.se/
Copyright
The publishers will keep this document online on the Internet - or its possible
replacement - for a considerable time from the date of publication barring
exceptional circumstances.
The online availability of the document implies a permanent permission for
anyone to read, to download, to print out single copies for your own use and to
use it unchanged for any non-commercial research and educational purpose.
Subsequent transfers of copyright cannot revoke this permission. All other uses
of the document are conditional on the consent of the copyright owner. The
publisher has taken technical and administrative measures to assure authenticity,
security and accessibility.
According to intellectual property law the author has the right to be
mentioned when his/her work is accessed as described above and to be protected
against infringement.
For additional information about the Linköping University Electronic Press
and its procedures for publication and for assurance of document integrity,
please refer to its WWW home page:
http://www.ep.liu.se/
Master’s Thesis 30 hp
Emma Persson
1 Introduction 5
2 Theory and background 6
2.1 Printed electronics . . . 6
2.2 ZnO . . . 7
2.2.1 The tetrapod shaped nanocrystal ZnO . . . 7
2.3 Schottky diodes . . . 9 2.3.1 Schottky junctions . . . 9 2.3.2 Ohmic contacts . . . 11 2.3.3 I-V characteristics . . . 13 2.3.4 C-V characteristics . . . 13 2.4 Rectifiers . . . 14
2.4.1 Half wave rectifiers . . . 14
2.4.2 Full wave rectifiers . . . 15
3 Experimental details 17 3.1 ZnO mix . . . 17
3.2 Electrode materials . . . 18
3.3 Printing . . . 18
3.4 Measurements . . . 19
4 Results and discussion 20 4.1 The effect of different ZnO concentration . . . 20
4.2 Screening of suitable electrode materials. . . 22
4.3 Adhesion problems . . . 25
4.4 The effect of different UV-doses with and without annealing . . . 25
4.5 The effect on saturation current with a thicker mesh . . . 27
4.6 The effect on forward current with surface treatment . . . 27
4.7 Frequency measurements . . . 31
4.8 Physical characterization . . . 35
5 Conclusions and future work 37 5.1 Goals . . . 38
2.1 a) a screen printed image of Marilyn Monroe by Andy Warhol. b) A mesh with a
finished pattern . . . 6
2.2 Images showing ZnO the difference in electron transportation for a) nanorods and b) nanotetrapods . . . 7
2.3 Isolated metal and n-type semiconductor . . . 9
2.4 Schottky junction in equlibrium . . . 10
2.5 A schottky junction in forward bias and reverse bias. a) Forward bias. b) Reversed bias. . . 10
2.6 Ohmic contacts a) Metal work function close to or smaller than the electron affinity of the semiconductor. b) High doping. . . 11
2.7 Surface rougness which limits surface-to-surface contact. The more conctact points, the lower conctact resistance. . . 11
2.8 Images showing ZnO nanotetrapods ability to pierce materials and increase the amount of contactpoints. a) Increasing the contact points b) Piercing of Teflon, taken from ref. [15] . . . 12
2.9 I-V characteristics of a typical Schottky diode . . . 13
2.10 C-V characteristics for a Schottky diode . . . 13
2.11 A half wave rectifier . . . 14
2.12 A capacitor has been added to a half wave rectifier decreasing the ripples and increasing the effective DC voltage . . . 15
2.13 a) a full wave rectifier and the current in b) the positive half cycle and c) the negative half cycle. . . 15
2.14 A capacitor has been added to a full wave rectifier to decrease the ripples and increase the effective DC voltage . . . 16
3.1 a) A Roku Print prepared and ready for printing and b) a fully printed sample with Carbon as top electrode and Silver as bottom electrode . . . 18
3.2 A rectifier bridge with Aluminum as ohmic contact and Silver as Schottky contact 19 3.3 How to probe the diode . . . 19
4.1 Profilometer measurements (conf20x magnification) of mixes in table 4.1. a) EP010, b) EP011, c) EP012, and d) EP013. . . 21 4.2 Picture of Mix 5 taken with an optical microscope at 100x magnification. The
4.4 Ag/binder 1/Ag printed and measured to see how much leakage the binder is responsible for. . . 23 4.5 I-V measurements of EP030. A Schottky diode with a bottom electrode of Al
(ohmic), 3,68 percent ZnO in binder 1 and a top electrode of Ag1 (schottky) annealed in 120 °C for 15 min) . . . 24 4.6 Diode A4 from sample EP079 with a rectification ratio of 2,6*10ˆ5 at +/- 2V . . 27 4.7 Sample EP110 diode A7. It has a high rectification ratio at 2 V but the forward
current is too low to reach the target of 0.1 mA at 2 V. . . 28 4.8 Sample EP110 diode B4. It has a lower rectification ratio at 2 V (7000 times)
with a bit higher saturation current. The forward current is higher (30 kOhm) and is almost reaching the goal of 20 kOhm. . . 29 4.9 Sample EP114 diode A9. It has a high rectification ratio at 2 V (100 000 times)
with a low saturation current but the forward current is only in the uA range. . . 29 4.10 Diode EP112A4 measured 1 h after print and 3 days after print. Note that the
saturation current has decreased after 3 days. . . 30 4.11 Frequency measurements of diode EP110A7 at a) 100 Hz, b) 200 Hz, c) 1 kHz, d)
10 kHz, e) 100 kHz and f) 1 MHz. . . 31 4.12 Frequency measurements of diode EP110B4 at a) 100 Hz, b) 200 Hz, c) 1 kHz, d)
10 kHz, e) 100 kHz and f) 1 MHz. . . 32 4.13 Frequency measurements of diode EP119D7 at a) 1 kHz, b) 10 kHz, c) 100 kHz
and d) 1 MHz. . . 33 4.14 Frequency measurements of a rectifier bridge with sample nr EP132B1. The black
dotted line shows measurements from a commercial rectifying bridge. a) 100 Hz, b) 1 kHz, c) 10 kHz and d) 100 kHz. . . 34 4.15 Non-anneald EP033A9 a) before and b) after tape-test. Ag as top and bottom
electrode with only the binder without ZnO. . . 35 4.16 Non-anneald EP040A6 a) before and b) after tape-test. Ag as bottom electrode
and C as top electrode. . . 35 4.17 A SEM picture of a crack in the EP013 binder. EDX tells there is a lot of
mag-nesium in it . . . 36 4.18 a) An SEM picture of a ZnO nanotetrapod inside the EP012 binder, one can see
the legs point in different directions. 20k magnification, 20 kV. b) An SEM picture of a ZnO nanotetrapod inside the EP013 binder, one can see the sharp peak. . . 36
Abstract
The aim of this master thesis was to develop a process for fabricating Schottky diodes, using techniques that are suitable for cheap large volume mass production e.g. printing, with tetrapod structured ZnO as the semiconductor. Part of the work involved selecting suitable metals for ohmic and Schottky contact and identification of a binder that can be used for dispersion of the Zinc Oxide (ZnO). ZnO is a II-VI compound semiconductor with a wide band gap (3.4 eV). The Schottky diode is used as a rectifier. A rectifier serves the purpose to turn Alternating Current (AC) to Direct Current (DC). The Schottky diode should only conduct current in the forward direction, in the reverse direction the current should be blocked. In this thesis printed diodes were used to construct different types of rectifiers for example half wave rectifiers and full wave rectifiers. Aside from electrical properties, adhesion properties have also been investigated. Ad-hesion was showed to depend on not only the choice of binder, but also UV-dose and annealing temperature. Aluminum and silver together with ZnO proved to be the best materials combina-tion with a rectificacombina-tion ratio up to 105− 106. Different sizes of Schottky diodes were printed and
the smaller diodes with an area of 0, 5x0, 5mm2 performed best as a half wave rectifiers while
Introduction
Thin Film Electronics ASA is a company that produce electronics with the printed electronics techniques. It is a Norweigan company with its headquarters in Oslo and product development in Linkoping. They were the first to commercialize printed rewritable memory and are now creating system products that will include memory, sensing, display and wireless communication that will be part of integrated systems making the way for electronic intelligence in applications where they have never been affordable before[1]. Some integrated systems need printable diodes that for example can convert a DC signal from an AC input. In this project nano tetrapods of Zinc Oxide are used as the semiconductor in a Schottky diode.
Zinc Oxide (ZnO) is a II-VI compound semiconductor with a wide band gap (3.4 eV) [2]. The tetrapod ZnO nanostructure material that are used is one of several ZnO nanostructures and is commercially manufactured and is called the "Pana-Tetra" [3]. One of the advantages with the tetrapod structure is that the tetrapods may spontaneously orientate with one of the four arms directed normal to the substrate [4].
Metal contacts are required for almost any electrical device ZnO can be used for, but there is a lack of information about printable Schottky diodes and Schottky diodes based on ZnO nano structures even though studies about Schottky barriers of metals on ZnO began in the mid-1960s [5]. A better understanding of how the ZnO surfaces, ZnO-metal interfaces and the processes involved during contact formation affect the electronic properties is needed [5].
The aim of this master thesis is to develop a process for fabricating printable Schottky diodes based upon ZnO materials for Thin Film Electronics. The work is focused on formulating screen printable ZnO inks using commercially available ZnO particles with a tetrapod structure and to mix these formulations with commercially available pre-formulated binders used in the printed electronic industry. The effect of different surface preparations is a relatively unexplored area and hence it’s hard to predict which methods are suitable. Surface preparations may be of impor-tance, but for Roll-to-Roll production in air at ambient pressure the number of preparation steps are limited. therefore this project is focused on the use of different contact materials, binders, junction areas and junction thicknesses, annealing temperatures and UV-doses.
Theory and background
2.1
Printed electronics
When people hear the word "Screen printing" they might think of Andy Warhol, the American artist responsible for pop art paintings like Marilyn. Others might think of T-shirt prints or something totally different.
(a) (b)
Figure 2.1: a) a screen printed image of Marilyn Monroe by Andy Warhol. b) A mesh with a finished pattern
Today this printing technique has evolved and is not only used in the art industry, but in several other areas such as printed electronics. Printing of functional materials such as conductive polymers, inorganic semiconductors etc. can be used as a low-cost and flexible technique that in some aspects can compete with traditional electronics for example cost per tag [6], [7]. The basic technique itself is the same as before. To make a pattern a woven mesh is used and then filled with a layer of photo emulsion and dried. The pattern is placed on the screen and then exposed to light. The emulsion rigidifies and binds to the woven mesh except for the area where the pattern covers the emulsion [8]. When the mesh is finished one can print the pattern on a substrate by pressing ink against the mesh. The most important rule in screen printing is to adjust the mesh to the ink instead of adjusting the ink to the mesh. The ink should only be
2.2
ZnO
Zinc Oxide (ZnO) is a II-VI compound semiconductor with a direct wide band gap (3.37 eV). It has a stable wurtzite structure and a lattice spacing a=0.325 nm and c=0.521 nm [2]. Because of intrinsic defects such as O vacancies and Zn interstitials, ZnO is an n-type semiconductor by nature [10]. Because of its wide band gap and unique properties it is suitable for a broad range of applications such as UV-light emitters, chemical sensors and transparent electronics. Because of this ZnO is attracting more attention now than ever before [5]. ZnO is even suitable for space applications due to its high resistivity to high-energy radiation [11]. Several types of nano structured ZnO exist and can be easily synthesised such as nanohelixes, nanorings, nanosprings, nanorods, nanobelts and nanotetrapods [12]. In this project a ZnO powder of nanotetrapods mixed with an organic binder is used.
2.2.1
The tetrapod shaped nanocrystal ZnO
The tetrapod structured ZnO material is one of several ZnO nanostructures and has been com-mercially manufactured by Panasonic and is called the "Pana-Tetra".[3] The actual size of the tetrapods in this project is not in the nanometer scale but in the micrometer scale. Some of the physical properties of semiconductor materials undergo changes when one shrink the dimensions down to nanometer scales and the smaller tetrapods might behave differently than the ones used in this project [2].
There is little information about how the ZnO tetrapod differs from bulk ZnO and other struc-tures but one of the known advantages with the tetrapod structure is that the tetrapods may spontaneously orientate with one of its four arms directed normal to the substrate [4]. The nan-otetrapod consist of four nanorods joined at tetrahedral angles to a central core. This gives it the advantage of higher vertical conduction as shown in figure 2.8b. The nanorods usually lie down horisontally on its substrate and can not use the nanorods one dimensional electron transport in the vertical direction in comparison to the tetrapods that can use the one dimensional electron transport in a direction perpendicular to the conductive substrate giving the electrons a shorter distance to travel [13], [14].
(a) (b)
Figure 2.2: Images showing ZnO the difference in electron transportation for a) nanorods and b) nanotetrapods
ZnO may also be used to modify adhesion properties. At the Zoological Institute of Kiel University ZnO tetrapods have succesfully been used to stick polymer Teflon to Silicone [15]. Both Teflon and Silicone are non-sticky materials that are joint together by using the tetrapods as linkers as the needle shaped nanoparticles. A SEM picture of this can be seen in figure 2.8b. Wei Chen et al achieved high performance solar energy conversion photoelectrodes based on
ZnO nanotetrapods [16]. They showed that ZnO is superior for electron transport and collection when assembled into a network.
The properties of the commercially manufactured "Pana-Tetra" is given in table 2.1. Table 2.1: Properties of the ZnO tetrapods used in this project
Material name Zinc Oxide
Chemical formula ZnO
Structure Single crystal (Needle shape)
Shape Tetrapod Shape
Average length of leg ∼ 10 − 20µm
Specific gravity 5.78
Relative density ∼ 0.1
Melting point under pressure 2000◦C
Sublimation point 1720◦C
Specific heat 0.1248cal/g · deg
Thermal conductivity 25.3W/m · K
Thermal expansion coefficient 3.18 · 10−6/K
Refractive index 1.9-2.0
Electricity induction (2.4 · 1010Hz) ǫ = 8.5
2.3
Schottky diodes
Metal contacts are required for almost any electrical device ZnO can be used for, but there is a lack of information about printable Schottky diodes and Schottky diodes based on ZnO even though studies about Schottky barriers of metals on ZnO began in the mid-1960s. A better understanding of how the ZnO surfaces, ZnO-metal interfaces and how the processes involved during contact formation affect the electronic properties is needed [5]. The effects of different surface preparations is a relatively untouched area, consequently there is not enough information about the properties of the contact between ZnO and different metals. Two forms of contacts can be formed between a metal and a semiconductor, Schottky barriers and the ohmic contacts. Schottky barriers act as rectifiers that allows current in only one direction and blocks it in the other direction while ohmic contact is a symmetric type of contact that conducts in both directions. The goal of this project is to produce a printed Schottky diode with an series resistance less than 20 kOhm, a saturation current larger than 10−15 and a zero bias junction capacitance
less than 100 pF. Different electrode materials, binders and the effect of the thickness of the binder and semiconductor layer are studied.
2.3.1
Schottky junctions
Schottky barriers act as rectifiers and are suitable for high-speed applications due to their high switching speed. Theoretically a Schottky junction is formed if the metal work function is larger than the n-type semiconductor work function, for a p-type semiconductor it’s the other way around. Consider a metal and a n-type semiconductor at thermal equlibrium with the energy states shown in figure 2.3. The metal has a workfunction φm and the semiconductor has a
workfunction φs. The Fermi level for the metal is EF mand for the semiconductor EF s[17],[18].
EV ac EF m EV EC EF s qφm Eg qχs qφ s
Figure 2.3: Isolated metal and n-type semiconductor
When the metal and the semiconductor are brought to contact the Fermi levels align and a band bending in the energy diagram is observed (figure 2.4). By calculating the difference in work function across the interface the junction built-in-potential, φi, for the semiconductor is
obtained as seen in figure 2.4.
φi= φm− φs
The barrier potential φb can be calculated as
φb= φm− χs
EV ac EV EC EF qφm Eg qχs qφs W qφb qφi
Figure 2.4: Schottky junction in equlibrium
Applying a bias, VA, causes the Fermi levels of the semiconductor and the metal to split up
with the energy difference of qVAeV . In a forward bias the band bending is less significant,
the depletion region is reduced and a large current flows. In reverse bias on the other hand the depletion region width is increased and the current is consequently very small. This means that if an alternating voltage is applied across the junction the charges can flow in only one direction and therefore the current is rectified. This is explained more thoroughly in section 2.4.
Electronf low − + V EV EC EF s qφb qVA q(φi− VA) (a) + − V EV EC EF s qφb |qVA| q(φi− VA) (b)
Figure 2.5: A schottky junction in forward bias and reverse bias. a) Forward bias. b) Reversed bias.
ZnO has an electron affinity of qχs ≈ 4.2eV [5]. Metals with workfunctions a bit higher than qφm= 4.2eV are candidates to form Schottky contact with ZnO. In table 2.2 materials with work functions close to the desirable value are listed.
Table 2.2: Suitable metals for ohmic contacts on ZnO
Material Work function qφm(eV )
Ag 4.52 - 4.74 [19] PEDOT:PSS 5.1 [20]
2.3.2
Ohmic contacts
The ohmic contact is a linear type of contact in which voltage- current relation follows the Ohm’s law [23]. Ideally, the ohmic contact should have a linear I-V characteristic for both forward and reversed bias, it does not perturb device performance significantly due to a small voltage drop across the contact and the contact resistance. An ohmic contact is formed between an n-type semiconductor and a metal if φmis close to or smaller than the semiconductors electron affinity,
χs, the barrier height is reduced so that the electrons can flow in both directions, see figure 2.6a.
electronf low EV ac EV EC EF s qφm qχs qφb (a) tunneling EV EC EF s qφb (b)
Figure 2.6: Ohmic contacts a) Metal work function close to or smaller than the electron affinity of the semiconductor. b) High doping.
High resistance in metal-semiconductor ohmic contacts are often caused by contact failure or high thermal stress, which leads to a major loss of device performance [23]. (See figure 2.7)
Figure 2.7: Surface rougness which limits surface-to-surface contact. The more conctact points, the lower conctact resistance.
Our hypothesis is that the nanotetrapod ZnO pierce the contact material, figure 2.8a, and form more contact points resulting in a low contact resistance, as seen in figure 2.8a. Moreover, we believe that this leads to a junction with good mechanical properties and contact strength.
(a) (b)
Figure 2.8: Images showing ZnO nanotetrapods ability to pierce materials and increase the amount of contactpoints. a) Increasing the contact points b) Piercing of Teflon, taken from ref. [15]
As discussed before one of the contacts in a diode should be ohmic. The desirable properties of such contact can be summarized as below:
• The metal work function should be close to or smaller than the semiconductor electron affinity φm.χs.
• The surfaces should have as many contact points as possible in order to have a smaller contact resistance
A survey of the literature presented in the field does not give a clear answer of which metals are suitable to form an ohmic contacts with ZnO. ZnO has an electron affinity of qχs≈ 4.2eV [5]. Metals with workfunctions qφm .4.2eV are candidates to form ohmic contact with ZnO. The binder we are using with the ZnO powder might change the electron affinity but we assume the work function to be constant. In table 2.1 metals with work functions close to the desirable value are listed.
Table 2.3: Suitable metals for ohmic contacts on ZnO
Metal Work function qφm(eV ) C 5 [19] Sn 4.42 [19] Al 4.06-4.41 [19]
2.3.3
I-V characteristics
According to the Schockley diode equation the current running through the diode can be de-scribed as [21]:
I = Isat(e
q(VA−φi)
ηkT − 1)
where Isatis the reverse saturation current, φi is the barrier height, VA is the applied voltage, q
is the elementary charge, K is the Boltzmann constant, T is the diode temperature and η is the ideality factor, 1 ≤ η ≤ 2.
In reversed bias the applied voltage VAis negative leading to a negligably small current Isat.
The I-V characteristic of a typical Schottky diode is shown in figure 2.9.
I
V
Figure 2.9: I-V characteristics of a typical Schottky diode
2.3.4
C-V characteristics
The Schottky junction act as a parallel plate capacitor due to the depletion region width between the metal and the doped semiconductor that acts as an insulator. The depletion region decrease with forward bias and increase with reversed bias. The junction capacitance-voltage relation can be expressed as [22]:
CJ(VA) = CJ(0) (1−VAφi)1/2
V C
2.4
Rectifiers
The Schottky diode is used as a rectifier. A rectifier serves the purpose of turning Alternating Current (AC) to Direct Current (DC). The Schottky diode should only conduct current in the forward direction from its anode to its cathode. In the revers direction, from cathode to anode, the current is blocked. In an n-type diode the metal part acts as the anode and the semiconductor acts as the cathode. In a p-type diode the metal part acts as the cathode and the semiconductor as the anode. Diodes can be used to make different types of rectifier circuits for example half wave rectifiers and full wave rectifiers. The half wave rectifier has only one diode and the full wave rectifier can have two diodes, or four diodes as in a diode bridge.
2.4.1
Half wave rectifiers
The simplest rectifier is the half wave rectifier which contains only one diode. In the ideal case, when an ac voltage is applied across the diode the current is blocked during the negative half cycles but in the positive half cycles it lets current through making it unidirectional, i.e. it turns AC to DC.
Figure 2.11: A half wave rectifier
The load resistor in figure 2.11 gives a proportional relationship, U = RI, between the current and the voltage across it in the forward direction. The voltage measured in the forward direction is equal to the supply voltage, VS, minus the voltage drop through the diode in forward
bias, VF.
Vout= VS− VF
The effective DC voltage, i.e. the mean value of the output voltage, can be written as:
VDC= Vmax/π
where Vmax is the maximum value of the output voltage. The effective DC current and the
forward bias the capacitor is charged and during reversed bias the capacitor is slowly discharged until the next half cycle of forward bias, keeping the output voltage more stable (see figure 2.12). This gives a larger effective DC voltage and smaller ripples. But for a half wave rectifier this capacitor needs to be relatively large because of the large ripples and low ripple frequency due to the cancelled out half cycles, therefore half wave rectifiers are used in low-power applications. For high-power applications a full wave rectifier is more suitable.
Figure 2.12: A capacitor has been added to a half wave rectifier decreasing the ripples and increasing the effective DC voltage
2.4.2
Full wave rectifiers
With four diodes one can create a diode bridge. This kind of rectifiers passes the full wave from input to output but they invert the negative half cycle. During the positive half cycles current goes through diode D1 and D2 while diode D3 and D4 are off (figure 2.13b). During the negative half cycles current goes through D3 and D4 while diode D1 and D2 are off instead (figure 2.13c)[25].
(a)
(b) (c)
Figure 2.13: a) a full wave rectifier and the current in b) the positive half cycle and c) the negative half cycle.
voltage, the DC output of a full wave rectifier is twice as large as a half wave rectifier. But due to the fact that the current passes two diodes in series the voltage drop is also twice as large compared to half wave rectifier case. The voltage measured in the forward direction should be the same as the supply voltage, VS, minus the forward voltage drop in the diodes, VF, that is
typically 0.7 V for one Schottky diode.
Vout= VS− 2VF
The effective DC voltage, i.e. the mean value of the output voltage, can now be written as:
VDC= 2Vmax/π
The effective DC current and output power can be calculated in the same way as for the half wave rectifier. A capacitor can be added in parallel with the resistive load to diminish the ripples and achieve a higher dc output (see figure 2.14).
Figure 2.14: A capacitor has been added to a full wave rectifier to decrease the ripples and increase the effective DC voltage
Experimental details
In this project several parameters are optimized by examining the effect of different ZnO con-centrations, mesh sizes, electrode materials, different UV-doses, annealing time and annealing temperatures.
3.1
ZnO mix
The ZnO nanotetrapods were delivered as a powder. Since it is not possible to print a powder, it had to be mixed in a binder. Seven different UV-curable binder materials were examined. Different concentrations of ZnO in different binders were prepared as seen in table 3.1.
Table 3.1: List of the different mixes used in the project
Mix Binder ZnO mass conc.
#1 Binder 1 1,02% #2 Binder 2 0,88% #3 Binder 1 3,68% #4 Binder 3 3,65% #5 Binder 1 3,70% #6 Binder 4 1,70% #7 Binder 5 3,10% #8 Binder 6 3,70% #9 Binder 7 5,20% #10 Binder 1 5,70% #11 Binder 1 + retarder 9,34%
3.2
Electrode materials
Several electrode materials were used as listed in table 3.2 to find the best combination of Schottky and ohmic contact for the diode. Two different types of Ag ink were used which are referred as Ag ink (1) and Ag ink (2) in the table.
Table 3.2: Combinations of electrode materials that were tested
Ohmic contact Schottky contact
C Ag ink (1) C PEDOT:PSS C Ag2 Al Ag ink (2) Al PEDOT:PSS Al Ag2
3.3
Printing
The printing equipment used was a Roku Print screen printing machine RP 2.2. This equipment lets the user save settings like positions and speed in a program that makes the printing easier to replicate. A velocity of 25% was used for all samples.
The Schottky diodes were printed on a PEN substrate in a pattern where ZnO is sandwiched between a top electrode and a bottom electrode. To do this three meshes were needed. One for the bottom electrode, one for the ZnO layer and one for the top electrode. During the project the size of the bottom and top electrode meshes were 140-34 and 100-34 respectively for all samples. Two different sizes were used for the ZnO layer. Most of the samples were printed with 120-34 size but the mesh size 61-64 was tested as well. The Al electrodes were patterned in another way since screen printing Al is not possible. A pattern of four different surface areas, 0.5x0.5mm2,
1x1mm2, 2x2mm2 and 4x4mm2 was designed as shown in figure 3.1b. Since the diode partly
works as a parallell-plate capacitor the contact area affects the performance of the diode.
Figure 3.2: A rectifier bridge with Aluminum as ohmic contact and Silver as Schottky contact
3.4
Measurements
The I-V-measurements were done with a Keithley 4200-SCS. Three to five measurements were done before saving the plots to make sure that the I-V characteristic of the device was stable.
Figure 3.3: How to probe the diode
The half- and full-wave measurements were performed using an input AC voltage from an arbitrary waveform generator (National Instruments PXI-5412) and the output voltage was read with a digitizer (National Instruments PXI-5114) with a load resistance of 1 MOhm. The thick-ness and surface roughthick-ness of the samples were characterized with a dektak profilometer.
Results and discussion
4.1
The effect of different ZnO concentration
One advantage of using a ZnO powder is that the concentration of ZnO in the binder does not have to be fixed. This gives a possibility to change the electrical properties while keeping the size of the diode fixed. The concentration is an important parameter that affects the electrical properties and viscosity of the mix. A low concentration of ZnO gives lower currents and allows the diode to work at higher voltages. A high concentration of ZnO gives a rougher surface, higher currents but does not allow high voltages. Screen printing inks with ZnO concentrations of 0,88%, 1% and 3% were prepared and printed on Si wafers and characterized with a Dektak profilometer (Figure 4.1). At 6% ZnO binder 1 the viscosity becomes to high for screen printing, therefore the concentration has to be lower than 6% for binder 1, if no retarder is added. The concentrations and sample names are listed in table 4.1.
Table 4.1: Samples with different ZnO concentrations prepared for roughness characterization
Sample Binder ZnO mass conc.
EP010 mix#1 1,02% EP011 mix#2 0,88% EP012 mix#3 3,68% EP013 Binder 1 0%
A significant difference in topography is observed between the different concentrations listed in table 4.1. The binder without ZnO (EP013) gives a smooth surface if compared with the mixes with 1% ZnO concentration (EP010 and EP011) that has peaks on the surface. In the sample with higher ZnO concentration (EP012) the peaks are more concentrated, proving that the ZnO concentration affects the surface topography. The ZnO nanotetrapod’s legs have a length of 10-20 um and the height of the standing pods is 18-37 um. therefore the thickness of the printed binder should be slightly less than 18-37 um for the pods to point out from the binder and make contact with the electrodes. Thickness measurements indicate that a screen size of 120-34 is
(a) (b)
(c) (d)
Figure 4.1: Profilometer measurements (conf20x magnification) of mixes in table 4.1. a) EP010, b) EP011, c) EP012, and d) EP013.
A characterization performed with an optical microscope (figure 4.2) shows that the ZnO tetrapods are in the um-scale and can easily be seen with a magnification of 100x due to the transparency of the binder.
Figure 4.2: Picture of Mix 5 taken with an optical microscope at 100x magnification. The ZnO tetrapods can easily be seen with a magnification of 100x.
4.2
Screening of suitable electrode materials.
The choice of materials for the ohmic and the Schottky contacts is one of the most important steps in this project. Empirical tests with different combinations of materials and binders show that Ag ink 1 as a Schottky contact and Al as an ohmic contact are suitable for fabrication of a Schottky diode with ZnO as a semiconductor. Carbon can also be used for an ohmic contact, but it is not ideal. Three samples were printed in the orders of C/ZnO/C, Ag/ZnO/Ag and Ag/only binder/Ag to examine the contact properties. In figure 4.3a the I-V curve of a sample with top and bottom carbon electrode is shown. The four different plots are measurements from the four different sizes, A, B, C and D. Higher currents can be expected from the largest size, D. The I-V curve of all the samples are symmetric, but not fully linear. This indicates that carbon is not a good choice to provide an ohmic contact with ZnO. In figure 4.3b the sample with the printed silver against silver is plotted. Three of the sizes (B, C and D) are plotted. These plots have a slightly less symmetrical appearance, probably due to a better contact in either of the electrodes. The current is low indicating that Ag ink 1 is working as a Schottky contact since a low leakage current is the goal. It is observed that the current is blocked differently in the forward and the reversed region and that the size of the diode current does not seem to scale with the surface area. The hysteresis is very large. The difference between the materials is later proved to be significant with controlled parameters. The measurement of one sample with a Ag/only binder 1/Ag print (figure 4.4) shows that the leakage current in binder 1 is in the pA-scale meaning that it is the ZnO, and not the binder, that is responsible for almost all the leakage. The samples prepared with Ag ink 1 have a better performance compared to the ones prepared with PEDOT:PSS. Ag ink 2 have an extremely poor adhesion to the binder and PEDOT:PSS is too thin due to the screen size that is not suitable for the low viscosity of PEDOT:PSS.
(a)
(b)
Figure 4.3: Ohmic to ohmic vs. Schottky to Schottky contacts. a) C/ZnO/C and b) Ag/ZnO/Ag.
−5 0 5 10−14 10−13 10−12 10−11 EP033 Voltage [V] Current [A]
Figure 4.4: Ag/binder 1/Ag printed and measured to see how much leakage the binder is re-sponsible for.
A sample (EP030, see fig 4.5) with Al instead of C as a bottom electrode is printed to increase the current going through the diode. This sample conducts more current which is the purpose of the ohmic contact, but the leakage current is very high as seen in the negative voltage range in
figure 4.5. The four different plots are measurements from the four different sizes, A, B, C and D. The largest size (D) has the highest leakage and forward current as expected. The leakage has to do with the Schottky contact and controlled parameters can reduce the leakage significantly. Carbon has a high resistivity and the work function does not fit with the theoretical value as it does for Al. This makes Al a better ohmic contact as earlier expected. A decision to continue with Aluminum as an ohmic contact and Ag ink1 as a Schottky contact is made according to the experimental results from examining different electrodes.
Figure 4.5: I-V measurements of EP030. A Schottky diode with a bottom electrode of Al (ohmic), 3,68 percent ZnO in binder 1 and a top electrode of Ag1 (schottky) annealed in 120 °C for 15 min)
4.3
Adhesion problems
Adhesion problems are observed in all samples from the first batches. The mix does not stick to the substrate and the top electrodes crumble like sponges. A solution to the substrate adhesion is to anneal binder 1 for 10 min at 120 °C in after the UV treatment and before printing the top electrode. This is observed when printing mix #4 and mix #5 to test the annealing of binders prior to the printing of the top electrode. Humidity and UV-dose have to be controlled to solve the the adhesion problem of the top electrode.
Binder 4 shows a poor adhesion to Ag and Al electrodes, therefore it was discarded. Binder 6 has a good adhesion but the use of this material in the ZnO mix results in a lot of pinholes in the top electrodes. Despite the pinholes the two binders, binder 6 and binder 1 seem promising to continue with, by optimizing and controlling the UV dose.
4.4
The effect of different UV-doses with and without
an-nealing
The UV-dose the binder is exposed for affect the adhesion between binder and the substrate and electrodes. Deionization of the binder before printing the top electrode improves the adhesion of the Ag ink 1 to the binder.
The print trials continues with samples (EP060-EP063) with binder 6, mix #8, and Ag ink 1 as top electrode (table 4.2). The sample with the lowest UV-dose seems to have the best adhesion of Ag ink 1 on top of the binder, but the binder have poor adhesion to the substrate which makes both binder and top electrode fall of together.
Table 4.2: Prints to test UV-dose importance. All samples are printed with mix #8, mesh size 120-34 and Ag ink 1 as top electrode.
UV EP060 120W 580 Jx1 (+ 20 s deionization) EP062 120W 580 Jx1 EP063 80W 430 Jx1 EP066 120W 530 Jx1 EP067 120W 530 Jx1 (+ 20 s deionization) EP068 120W 413 Jx1 EP069 120W 413 Jx1 (+ 20 s deionization)
When deionizing the surface of the binder after UV curing the surface of the top electrode seems to be smoother. For sample EP066-69 it is clear that deionization improves the adhesion for Ag ink 1 on the binder as observed before. The binder has poor adhesion to the substrate and it is possible that this is a result of too low UV-dose.
Table 4.3: Prints to test UV-dose with and without annealing. The samples are printed with mix #5 and mix #8 with mesh size 120-34 and Ag1 as top and bottom electrode
ZnO UV speed Anneal Rest after anneal Comments
EP071-1 mix #8 580 Jx2 10 min 120 °C 6 min
EP072-1 mix #8 580 Jx1 no anneal
EP073-1 mix #8 580 Jx1 10 min 120 °C 6 min
EP074-1 mix #8 740 Jx1 no anneal smooth surface
EP075-1 mix #5 580 Jx3 10 min 120 °C 5 min
EP076-1 mix #5 580 Jx2 10 min 120 °C 5 min
EP077-1 mix #5 1086 Jx2 10 min 120 °C 5 min
EP078-1 mix #5 1086 Jx1 10 min 120 °C 5 min smooth surface
The humidity during print is 40 % for samples EP071-1 - EP078-1. All samples with binder 1 mix #5 look good and adhere well to the substrate. The samples with binder 6 mix #8 prepared with a low UV-dose (EP076-1 and EP078-1) have a smooth surface without crumbles in the top electrode and adheres well to the substrate. No electrical measurements are done since the samples do not have any bottom electrodes.
Sample EP075-2 - EP078-2 are printed in the same way as EP0xx-1 but with an added Ag ink 1 bottom electrode. All samples crumbles just as they did in the beginning of the project and this might be due to lower humidity on the day of print.
The samples from the two mixes that seem to have the best adhesion is EP078-1 and EP074-1. EP079 is printed with the same parameters as for EP078-1 but with Al bottom instead of Ag ink 1 and EP080 is processed in the same way as EP074-1 also using an Al bottom instead of the Ag ink (see table 4.4). The Al bottom is added to enable electrical characterization. Table 4.4: Prints with the same parameters as EP078-1 and EP074-1 with Al bottom instead of Ag ink 1
EP079 EP078-1 with Al bottom instead of Ag
EP080 EP074-1 with Al bottom instead of Ag
During processing the humidity is 40 %. Some of the samples have poor adhesion to the Al electrodes but overall these samples have a good smooth appearance. The rectification ratio of sample EP079 is measured (see fig 4.6) and relative to the first samples (fig 4.5) for example) a good rectification is observed. The rectification ratio is ca. 104 for all diodes at +/- 2 V but
for some of the diodes the rectification ratio gets as high as 105. The leakage current is low but
unstable (10−10− 10−8) and the forward current is not stable either. In order to control (achieve
a better stability) the forward current and saturation current attempts are made to surface treat Al.
−3 −2 −1 0 1 2 3 10−10 10−5 EP079A4 Voltage [V] Current [A] rect ratio=2.6e+05 mA at 2V=0.013
Figure 4.6: Diode A4 from sample EP079 with a rectification ratio of 2,6*10ˆ5 at +/- 2V A low UV-dose gives good results with binder 1. Due to very poor adhesion between binder 6 and Al in sample EP080 the binder 6 track is terminated. The samples with mix #5 is now having smooth surfaces without crumbled top electrodes and a good adhesion but almost all of the samples have electrical short circuits. It might have to do with mix #5 at this point is a few weeks old. A new mix with the same binder is made to continue with, mix #10.
4.5
The effect on saturation current with a thicker mesh
A low saturation current is a key performance indicator for the Schottky diode. By varying the UV-dose and mesh thickness the saturation current can be decreased. Several samples printed using mix #10 with different UV-doses (single exposure with 310 J to 740 J (80W)), with 10 min annealing in 120 °C and the same canvas thickness as before, 120-34, were examined. Single dose exposure of 580 J (120W) gives the best measurement data but the difference from the other samples is not significant. A similar test with a thicker canvas (61-64) in 65% humidity shows that the thickness of the binder affects the saturation current too. Significantly lower saturation current are observed for thicker canvas, but the forward current is also affected and the rectification ratio is constant, which implies that the difference is that the amount of contact points was decreased for thicker prints. The number of electrical shorts is lower for thicker prints, which is an advantage that likely can be explained by less pinholes.
4.6
The effect on forward current with surface treatment
A good diode does not only need a low saturation current, the forward current needs to be high as well. To increase the forward current the ohmic contact needs to have a low ohmic resistance. Al oxidizes in air and give a pinned layer to the semiconductor, increasing the ohmic resistance. With a quick surface treatment with Acetone, IPA and some distilled water, prior to printing of the ZnO layer, the forward current increase slightly. The two different thicknesses were used in the same sample batch, with mix #10, one without annealing and one annealed for 10 min in 120 °C. The two thicknesses show almost the same rectification ratio but the I-V plots have an offset of two decades. The I-V curve of a diode with the surface area of (0.5x0.5 (mm2), printed
with frame with the mesh size of 120-43 is illustrated in figure 4.7. The rectification factor at 2 V is 6.9 ∗ 104, the saturation current is the tenth of nA-range at -2 V and the forward current
is slightly above 10 uA at 2V. The target is an even higher forward current (0.1 mA at 2 V) to reach 20 kOhm resistance of the diode.
−3 −2 −1 0 1 2 3 10−12 10−10 10−8 10−6 10−4 EP110A7 Voltage [V] Current [A] rect ratio=6.9e+04
Figure 4.7: Sample EP110 diode A7. It has a high rectification ratio at 2 V but the forward current is too low to reach the target of 0.1 mA at 2 V.
In figure 4.8 the next smallest size, B (1x1 mm2), from the same sample show a less stable
saturation current and a lower rectification ratio (1.4 ∗ 104). The forward current is higher (30
−3 −2 −1 0 1 2 3 10−12 10−10 10−8 10−6 10−4 EP110B4 Voltage [V] Current [A] rect ratio=1.4e+04
Figure 4.8: Sample EP110 diode B4. It has a lower rectification ratio at 2 V (7000 times) with a bit higher saturation current. The forward current is higher (30 kOhm) and is almost reaching the goal of 20 kOhm.
A sample with the thicker mesh (61-64) but surface treated and annealed as sample EP110 shows excellent diode characteristics (see figure 4.9). The saturation current is as low as pA at -2 V and the forward current is almost reaching the uA level at 2 V giving it a rectification ratio of (1.3 ∗ 105). The forward current is too low to reach the goal, but with an offset of slightly more than two decades this diode would be reaching the goal of 20 kOhm and still having a saturation current under 1 nA.
−3 −2 −1 0 1 2 3 10−12 10−10 10−8 10−6 EP114A9 Voltage [V] Current [A] rect ratio=1.3e+05
Figure 4.9: Sample EP114 diode A9. It has a high rectification ratio at 2 V (100 000 times) with a low saturation current but the forward current is only in the uA range.
The analysis of the electric characteristics of devices prepared using mix 1-10 shows the necessity to increase forward current, yield and stability. The increase in forward current is likely governed by increasing concentration of the ZnO. However increasing ZnO concentration is not trivial since it increases viscosity. Therefore additional retarder was added to lower viscosity and possibly increase the levelling which could reduce loss yield caused by various sorts of print defects.
A new mix of binder 1 with a higher concentration of ZnO and an extra retarder was printed using two different canvases. The diodes were printed in the same way as the diodes in figure 4.7 and 4.9. Surprisingly none of the diodes are shorted, probably due to the retarder helping the binder not to crack when cured and annealed. A problem with these samples are that since the concentrations of ZnO is too high leading to higher currents going through the diode the diodes are shorted during measurements with a high voltage range.
Some samples show an aging effect that is the saturation current decreases during few days after printing. This effect, illustrated in figure 4.10, might have affected decisions on which samples were the best since some samples are measured the same day as they are printed and some of them are measured a couple of days later. Most of the later samples are measured after a couple of days. The red plot in figure 4.10 corresponds to a measurement 1 h after the print and the blue plot correspond to a measurement 3 days after print. The saturation current is lower but the forward current is unaffected by time.
Figure 4.10: Diode EP112A4 measured 1 h after print and 3 days after print. Note that the saturation current has decreased after 3 days.
4.7
Frequency measurements
Frequency measurements are carried out on some of the best diodes of three different sizes to see how the saturation current and forward current affect the rectifying behaviour. In diode EP110A7 the saturation current is really low and this can be observed in the frequency measurements (figure 4.11)as well as the I-V measurements in figure 4.7. At 100 Hz and 200 Hz the output current has an almost ideal appearance. The reversed current is almost blocked and the voltage drop is ca. 0.7 V as expected. At 1 kHz there is not enough time for the diode to discharge during the negative half cycle. When the frequency gets higher this behaviour is more pronounced and at 1 MHz the output voltage is almost a stable 2 V signal. At 1 kHz the voltage drop increased leading to a higher effective voltage.
Type Al Treatment ZnO Anneal
EP110A7 Al/ZnO/Ag Aceton+IPA+dest.vatten 120-34 mix #10 (5,20%) 10 min (120 C)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 x 104 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 100 Hz 4V AC Input Output (a) 0 0.5 1 1.5 2 2.5 x 104 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 200 Hz 4V AC Input Output (b) 0 1000 2000 3000 4000 5000 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 1 kHz 4V AC Input Output (c) 0 50 100 150 200 250 300 350 400 450 500 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 10 kHz 4V AC Input Output (d) 0 5 10 15 20 25 30 35 40 45 50 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 100 kHz 4V AC Input Output (e) 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 1 MHz 4V AC Input Output (f)
Figure 4.11: Frequency measurements of diode EP110A7 at a) 100 Hz, b) 200 Hz, c) 1 kHz, d) 10 kHz, e) 100 kHz and f) 1 MHz.
In diode EP110B4 the leakage and the forward current are higher than the device with size A (0.5x0.5mm2) from the same sample. The mean value of the output voltage remains constant
until 10 kHz but at 100 kHz starts to decrease. At 10 kHz an extra high leakage current can be observed. It might be due to the high frequencies heat contribution giving thermal excitation. At 100 kHz the built in capacitance in the diode helps by not letting the reverse current go through but this leads to a low forward current as well. At 1 MHz the output voltage still has a bit of a wave shape but even lowered.
Type Al Treatment ZnO Anneal
EP110B4 Al/ZnO/Ag Aceton+IPA+dest.wat 120-34 mix #10 (5,20%) 10 min (120 C)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 x 104 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 100 Hz 4V AC Input Output (a) 0 0.5 1 1.5 2 2.5 x 104 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 200 Hz 4V AC Input Output (b) 0 1000 2000 3000 4000 5000 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 1 kHz 4V AC Input Output (c) 0 50 100 150 200 250 300 350 400 450 500 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 10 kHz 4V AC Input Output (d) 0 5 10 15 20 25 30 35 40 45 50 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 100 kHz 4V AC Input Output (e) 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 −4 −3 −2 −1 0 1 2 3 4 Time [us] Voltage [V] 1 MHz 4V AC Input Output (f)
Figure 4.13 shows an example of a diode that is not optimal. The leakage is too high and it cannot operate in voltages higher than 2 V without burning off. Since the wave shape is almost following the input voltage even at 1 MHz it might work better at even higher frequencies but the measurement equipment could not measure at higher frequencies. It could be possible that diodes with this area and ZnO concentration could work better in series in a diode bridge rather than working alone.
Type Al Treatment ZnO Anneal
EP119D7 Al/ZnO/Ag Aceton+IPA+dest.wat 120-34 mix #11 (9,34%) 10 min (120 C)
0 1000 2000 3000 4000 5000 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.5 Time [us] Voltage [V] 1 kHz 2V AC Input Output (a) 0 50 100 150 200 250 300 350 400 450 500 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.5 Time [us] Voltage [V] 10 kHz 2V AC Input Output (b) 0 5 10 15 20 25 30 35 40 45 50 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.5 Time [us] Voltage [V] 100 kHz 2V AC Input Output (c) 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 −2.5 −2 −1.5 −1 −0.5 0 0.5 1 1.5 2 2.5 Time [us] Voltage [V] 1 MHz 2V AC Input Output (d)
Figure 4.13: Frequency measurements of diode EP119D7 at a) 1 kHz, b) 10 kHz, c) 100 kHz and d) 1 MHz.
Full wave rectification
In figure 4.14 measurements from one of the diode bridges that was printed can be seen. The bridges were printed with diodes of the two smallest sizes: 0,5x0,5 mm2 (A) and 1x1mm2 (B).
(a) (b)
(c) (d)
Figure 4.14: Frequency measurements of a rectifier bridge with sample nr EP132B1. The black dotted line shows measurements from a commercial rectifying bridge. a) 100 Hz, b) 1 kHz, c) 10 kHz and d) 100 kHz.
The smallest size did not work as good for the rectifier bridge as when used as a half wave rectifier. That might have been because of the high series resistance which is twice higher in a diode bridge compared to a half wave rectifier. The larger size (B) worked better in the rectifier bridge than as a half wave rectifier. When used alone the saturation current was to high but when connected in series the saturation current was decreased to almost zero as seen in figure 4.14a. In figure 4.14 a commercial diode bridge is compared to the sample EP132B1. The shape of EP132B1 signal is following the commercial bridge quite well except for a higher voltage drop. It can also be observed that in 100 kHz, in the negative half cycles the voltage drop is slightly lower than for the positive half cycles.
4.8
Physical characterization
When testing adhesion standard office-tape was used to see if the materials were stuck to each other or not. The two samples below, EP033 and EP040, are two of the first samples which show poor adhesion. The adhesion was good when an identical sample was annealed before tape-test.
(a) (b)
Figure 4.15: Non-anneald EP033A9 a) before and b) after tape-test. Ag as top and bottom electrode with only the binder without ZnO.
(a) (b)
Figure 4.16: Non-anneald EP040A6 a) before and b) after tape-test. Ag as bottom electrode and C as top electrode.
The samples EP012 and EP013 were gold plated using Chemical Vapour Deposition (CVD) ca. 10 Angstrom to be able to characterize them with a Scanning Electron Microscope (SEM). All that could be seen were cracks in the binder as seen in figure 4.17. When measuring with an Energy-dispersive X-ray spectroscopy in the cracks there were no trace of Zn but a lot of magnesium, silicon and oxygen. The magnesium probably derived from the binder and the silicon from the substrate since the crack opened up a hole down to the substrate. Cracks like these might have caused electrical shortages in some of the diodes. When increasing the electron beam voltage it was possible to see the tetrapods below the Au (see figure 4.18). An EDX measurement showed that there were both Zn and O in the tetrapods. The ZnO tetrapods do not point out of the binder as expected which might have to do with the screen being the right size in combination with the Au evaporation to cover all of the pods.
Figure 4.17: A SEM picture of a crack in the EP013 binder. EDX tells there is a lot of magnesium in it
(a) (b)
Figure 4.18: a) An SEM picture of a ZnO nanotetrapod inside the EP012 binder, one can see the legs point in different directions. 20k magnification, 20 kV. b) An SEM picture of a ZnO nanotetrapod inside the EP013 binder, one can see the sharp peak.
Conclusions and future work
Different sizes of Schottky diodes has been printed with good performance and a rectification ratio of 105− 106 for the best ones. These diodes have been proved to work both as half wave
rectifiers and full wave rectifiers. Identification of suitable materials for the Schottky and the ohmic contact was a fairly straightforward process. Aluminum and silver together with ZnO makes, with the printing technique, a relatively good Schottky diode. Carbon did not work as well as aluminum and this was expected due to the theoretical values of the work functions. It also had a high resistivity. The measurements of PEDOT:PSS was a bit unstable, possibly caused by the mesh making it too thin. The second silver-ink gave a poor printing result, probably because of the age of the ink.
The binders were a bigger obstacle than expected. They crumbled during the annealing of the top electrode and had a poor adhesion to the substrate. In the end binder 1 was chosen to be the binder in the final products since it adhered better and the I-V curves became more stable when annealing the binder before printing the top electrode. Short circuits, high saturation currents and low forward currents have been a large problem. Adding a retarder in the mix to lower the viscosity solved the yield problem. Lowering the UV-dose led to a reduction in saturation current. The experimental results showed that the best results came from a UV-dose that was low, but not too low. There are two possible explanations for these observations. The high UV-dose can affect ZnO polarity and affect the interface [23]. The high dose also lead to a higher temperature in the binder matrix, this often affects some mechanical properties of the binder such as an increase in the transition temperature.
The results of every diode vary a lot for most of the samples but usually the results are quite similar when the diodes are close to each other on the substrate. This indicates that it might have to do with the printing technique. Non-even pressure during the prints, air bubbles below the substrate or meshes that were poorly cleaned that might have caused this behaviour. The con-centration of ZnO in the binder is an important parameter. In mix 11 the concon-centration is really high ca 9.5 %. These diodes could not handle as high voltages as the other diodes could. Diodes prepared from mixes of lower ZnO concentration or with higher thickness could be operated at higher voltages without suffering shorts. It was not possible to find an optimal concentration or say which one was better than the others, since different concentrations and different sizes suit different purposes. This project suggests a method to tune the diode characteristics such as rectification, series resistance and the saturation current to fit different applications.
5.1
Goals
• Which is the best combination of the contact materials available in my project? Aluminum as an ohmic contact and the silver Ag1 as a Schottky contact proved to be the best contact material combination of the ones tested.
• Which is the best binders available in my project to use with the contact
materials and the ZnO?
After some problems with adhesion binder 1 proved to be the best binder. • How does the device stand outer strains?
This has not been tested in a large range, only in the production and measurement stage. The diodes’ saturation currents get lower when letting it rest in room temp for a couple of days. The binder does behave different depending on humidity when printing. In the measurements it was observed that the diodes wasn’t significantly light sensitive. When the currents reached 1 mA some of the diodes got shortened.
• Is the goal achieved? Is it a product that works good enough for commercial
use and Roll-to-Roll production?
The diodes were tested alone as a half wave rectifier and in series as a full wave rectifier. The smallest size performed well as a half-wave rectifier at 4V due to the low saturation current but had a too low forward current exceed the voltage drop in the next diode when connected in series. At higher voltages and ZnO concentrations this would probably not be a problem, which could be tested in future research. The next smallest size (B) did not work good alone due to high saturation currents but in series this was cancelled out and it performed quite good except for a high voltage drop. The diode is unfortunately not fully printable since the Aluminum bottom has been ordered and manufactured using a method that is not a printing method.
5.2
Future work
If performance requirements are set one could try to produce a diode with the right ZnO con-centrations and diode size to match a specific purpose. The forward currents of the small diodes are quite low but if one surface treats the aluminum with phosphoric acid, for a longer time than tried in the project, to remove the aluminum oxide the forward current might be higher. Also real capacitance vs. voltage measurements could be performed. The goal was to make a fully printable diode that was not reached, so new inks as a substitute for aluminum should be tried, for example tin-ink. Even though the diodes are not fully printable the diodes can still be mass produced with roll-to-roll production, but the process might have to be optimized for this type of production.
[1] www.thinfilm.no 2013-04-15
[2] Zinc Oxide Nanostructure: Synthesis and Properties Zhiyong Fan and Jia G. Lu, 22 April 2005.
[3] Whisker ultra-high strength single crystal published from SANGYO TOSHO
[4] ZnO tetrapod nano crystals Marcus C. Newton and Paul A. Warburton Materials today, 2007, 10, number 5.
[5] ZnO Schottky barriers and Ohmic contacts Leonard J. Brillson and Yicheng Lu Journal of
Applied Physics, 2011, 109, 121301.
[6] Printed electronics: The challenges involved in printing devices, interconnections and
con-tacts based on inorganic materials. J. Perelaer, P. J. Smith, D. Mager, D. Soltman, S. K.
Volkman, V Subramanian, J. G. Korvink and U. Schubert J. Mater. Chem,, 2010, 20, 8446-8453.
[7] http://www.lope-c.com, accessed 20th January 2010.
[8] http://www.instructables.com/id/Photo-emulsion-Screen-Printing/ , , , [9] How to be a great screen printer Professor Steven Abbot.
[10] Group III Impurity Doped Zinc Oxide Thin Films Prepared by RF Magnetron Sputtering Look, Reynolds Japanese Journal of Applied Physics, 1985, 24, L781
[11] Production and annealing of electron irridation damage in ZnO Look, Reynolds Applied
Physics Letter, 1999, 75, 811
[12] Zinc Oxide Bulk, Thin Films and Nanostructures: Processing, Properties and ... Chennupati Jagadish, Stephen Pearton.
[13] Photovoltqaic Devices Using Blends of Branched CdSe Nanoparticles and Conjugated
Poly-mers B. Sun, E. Marx and N. C. Greenham Nano Lett., 2003, 3, 961
[14] Hybrid Solar Cells with Prescribed Nanoscale Morphologies Based on Hyperbranched
Semi-conductor Nanocrystals I. Gur, N. A. Fromer, C.-P. Chen, A. G. Kanaras and A. P. Alivisatos Nano Lett, 2007, 7, 409
[15] Joining the un-joinable: Adhesion between low surface energy polymers using tetrapodal ZnO
linkers X. Jin, J. Strueben, L. Heepe, A. Kovalev, Y.K. Mishra, R. Adelung, S.N. Gorb, A.
[16] Branched ZnO nanostructures as building blocks of photoelectrodes for efficient solar energy
conversion Wei Chen, Yoingcai Qui and Shihe Yang Phys. Chem. Chem. Phys., 2012, 14,
10872-10881
[17] Semiconductor Device Physics and Simulation J. S. Yuan, Juin Jei Liou. [18] Introduction to solid state physics Charles Kittel.
[19] Handbook of Chemistry and Physics 66th edition CRS PRESS
[20] Conductivity, work function, and environmental stability of PEDOT:PSS thin films treated
with sorbitol A.M. Nardesa, M. Kemerinka, M.M. de Kokb, E. Vinkenc, K. Maturovaa,
R.A.J. Janssena
[21] http://ece-www.colorado.edu/ bart/book/ , , , [22] Schottky diodes Rick Cory, Skyworks Solutions Inc.
[23] A comprehensive review of ZnO materials and devices U. Ozgur, Ya. I. Alivov, C. Liu, A. Teke, M. A. Reshchikov, S. Dogan, V. Avrutin, S.-J. Cho and H. Morkoc Journal of Applied
Physics, 2005, 98, 041301.
[24] Selection of non-alloyed ohmic contacts for ZnO nanostructure based devices N. Koteeswara Reddy, q. Ahsanulhaq, J. H. Kim, M. Devika and Y. B. Hahn Nanotechnology, 2007, 18, 445710
[25] Basic Electronics Tutorials Wayne Stor
[26] Fundamentals of Zinc Oxide as a Semiconductor Anderson Janotti and Chris G Van de Walle Rep. Prog. Phys., 2009 72 126501.
