Amplification circuits based on electrochemical transistors

Title:

AMPLIFICATION CIRCUITS

BASED ON

ELECTROCHEMICAL TRANSISTORS

Master thesis

In

INFORMATION CODING

at

Linköping Institute of Technology

by

ZIA ULLAH KHAN

LiTH-ISY-EX--2008/4247--SE

Supervisor:

David Nilsson

Examiner:

Prof Robert Forchheimer

Presentation Date 2008-12-17

Publishing Date (Electronic version) 2009-01-15

Department and Division

Department of Electrical Engineering Information Coding

Language Type of Publication Title of thesis

Amplifier Circuits based on Electrochemical Transistors

English Master Thesis _Author

Number of pages Zia Ullah Khan

Series number

48 _{LiTH-ISY-EX--2008/4247--SE}

Abstract

Electrochemical Transistor (ECT) was reported by David Nilsson in 2002. Later, its dimensions were specified and then a simple SPICE model was developed.

The main purpose of this diploma work was to check the performance of electrochemical transistor in amplifier circuits. Simple amplifier circuits were simulated using SPICE model of ECT. Lateral and vertical structures of electrochemical transistors were patterned on Orgacon sheet (provided by AGFA), with various electrolytes (EG010, MS-HEC & MS-L). The characteristic curves and time responses of these ECTs were studied followed by checking them as an active component in the single transistor amplifier circuits. Screen Printed ECTs were also checked in amplifier circuits with the best available electrolyte. Behavior of self made and screen printed ECTs were compared on the basis of on-off ratio, slew rate, frequency response and gain. Screen printed ECTs showed promising results having less deterioration with time but till an input voltage’s frequency of 2 Hz only.

Abbreviations used:

ECT Electrochemical Transistor

LECT Lateral Electrochemical Transistor VECT Vertical Electrochemical Transistor

SP VECT Screen Printed Vertical Electrochemical Transistor SPICE Simulation Program with Integrated Circuit Emphasis FCC Forward Characteristics Curves

TCC Transfer Characteristics Curves OLED Organic Light Emitting Diode

EA Electron Affinity

IP Ionization Potential

Vgs Gate to Source Voltage Vds Drain to Source Voltage Ids Drain to Source Current Vp Pinch Off Voltage

gm Transconductance

HOMO Highest Occupied Molecular Orbital LUMO Least Unoccupied Molecular Orbital

Vs Supply Voltage

Rs Series Resistance

Abstract

Electrochemical Transistor (ECT) was reported by David Nilsson in 2002. Later, its dimensions were specified and then a simple SPICE model was developed.

The main purpose of this diploma work was to check the performance of electrochemical transistor in amplifier circuits. Simple amplifier circuits were simulated using SPICE model of ECT. Lateral and vertical structures of electrochemical transistors were patterned on Orgacon sheet (provided by AGFA), with various electrolytes (EG010, MS-HEC & MS-L). The characteristic curves and time responses of these ECTs were studied followed by checking them as an active component in the single transistor amplifier circuits. Screen Printed ECTs were also checked in amplifier circuits with the best available electrolyte. Behavior of self made and screen printed ECTs were compared on the basis of on-off ratio, slew rate, frequency response and gain. Screen printed ECTs showed promising results having less deterioration with time but till an input voltage’s frequency of 2 Hz only.

Mismatch with the simulation results and with the Shockley’s equation was also a finding after data analysis.

Acknowledgements

I would like to pay my sincere gratitude to David Nilsson for initiating the thesis topic within organic electronics. I appreciate his help throughout my diploma work as he made it extremely easy for me to work independently.

I am grateful to Prof Robert Forchheimer for guiding me with all his knowledge and experience. Special thanks to all the lab colleagues, creating a friendly atmosphere in clean room. Peter, Maria, Payman, Elias and all others were always there to help me in times of my distress.

I am also blessed with the prayers of my parents and acknowledge their support as well.

List of Contents

1: Introduction

1.1 Introduction to the goal of diploma work ………..….. 1

1.2 Flow chart of events ………..………..……… 1

2: Theoretical Background 2.1 Organic Materials 2.1.1 Conjugated Polymers ……….……..…. 3 2.1.1.1 Molecular Structure ……….…. 3 2.1.1.2 Band Structure ……….…. 6 2.1.2 Doping of a Polymer ………... 7

2.2 Charge Carriers and Charge Transport in a Polymer ………...………. 7

2.3 PEDOT:PSS as a Semiconductor ………..……. 9

2.4 Electrolytes ……….. ………..….… 9

3: Basic Structures 3.1 Structure 1 ………...…………... 10

3.2 Structure 2 ………... 11

3.3 Lateral Electrochemical Transistors ………. 11

3.4 Vertical Electrochemical Transistors ………... 12

4: Amplifiers 4.1 SPICE Model of Electrochemical Transistor ……….... 13

4.2 Various Amplifier Circuits ………..… 17

5: Experimental Part 5.1 Dimensions of an Electrochemical Transistors ………..… 20

5.2 Manufacturing Techniques ………..…. 20

5.2.1 Plotter Technique ………... 20

5.2.2 Screen Printing Technique ………..…. 21

5.3 Experimental Set up and Data Acquisition ………21

6: Results and Discussions 6.1 Lateral Electrochemical Transistors 6.1.1 With EG010 ……….. 22

6.1.2 With MS-L ………...24

6.2 Vertical Electrochemical Transistors 6.2.1 With EG010 ……….. 26

6.2.2 With MS-HEC ……… 28

6.2.3 With MS-L ………. 29

6.2.4 Frequency Response ………. 31

6.3 Screen Printed Electrochemical Transistors 6.3.1 With MS-L ……….. 32

6.3.2 Frequency Response ……….. 34

6.4 Linearity Measurement ……… 35

6.5 Slew Rate Measurement ………36

6.6 Comparison of self made and screen printed Electrochemical Transistors … ……… 37

6.7 Change in response for certain variables…… ………37

6.8 Set of Weird Observations ………..….40

1: Introduction

1.1 Introduction to the goal of diploma work

This diploma work was carried out at Acreo AB, using the clean room facilities. Lots of efforts are being put into the idea of producing cheaper electronic products on environmental friendly plastic substrate by printing techniques.

To bring this to reality, the basic component (Electrochemical Transistor) of an electronic circuit needs in depth studies and its behavior needs to be characterized. The Electrochemical Transistor was invented at Acreo and Linköping University few years back. Two of its geometries had been tested and its time response and characteristic curves showed promising results and it resembled to the conventional Field Effect Transistor (FET). Another diploma work was carried out to make a simulation model of this electrochemical transistor (ECT) in SPICE. A model of depletion mode FET was chosen as reference and was modified according to the output characteristic curves and time response data of the electrochemical transistor.

The diploma work carried out by me was a step further to verify the simulation results in real time circuits using self made ECTs. Lateral, Vertical and Screen Printed ECTs were used in various amplifier circuits with EG010, MS-HEC and MS-L as electrolytes.

1.2 Flow chart of events

A simple amplifier circuit was tested with the SPICE model of ECT. Parameter Analyzer was used to characterize these ECTs for their output and Transfer characteristic curves. The simulated amplifier circuit was tested with these ECTs and data was recorded in a PC. If the circuit responded satisfactorily, then its response was tested with defining certain circuit parameters. Other amplifier circuits were simulated using SPICE model and they were also tested in circuits. The first structure was of a lateral ECT and EG010 was used as electrolyte. Then the structure was changed to vertical with the same electrolyte. The electrolyte was then changed to MS-HEC and then to MS-L to get a faster ECT and hence better response from these amplifiers. Screen printed vertical design was also tried as an amplifier circuit element.

Simulation of a simple amplifier circuit Manufacturing of Electrochemical Transistors (ECTs) Checking their Transfer and Output characteristics

Checking the Transistors with

real resistors Is the amplifier working properly? No Yes Look for a better Electrolyte & Structure Simulate other amplifier circuits Checking the amplifier circuit for

certain variables

Verify those circuits with

real Transistors

Compiling the oddities of these

amplifier circuits

2. Theoretical Background

2.1 Organic Materials

Silicon Technology has reached ultimate heights and it is part of almost every electronic device in our daily life. Plastics were thought of being passive material only and found its use in insulation and packaging. In 1976, Heeger, MacDiarmed and Shirakawa came across an astonishing discovery when they witnessed metallic character in a mixture of polyacetylene (a polymer) and bromine[1]_{. From then}

on, organic electronics has emerged as an exciting new field for researchers in the last three decades. An organic material has been used as active semi conducting layer in the electronic devices i.e. organic light emitting diode emits light when current passes through organic semiconductor film and thin film transistors can behave like traditional silicon transistors.

It is called organic electronics because its basic component is carbon which is in abundance in all living organisms

Organic materials offer many advantages in comparison to inorganic materials.

• Organic materials are solution processable, so its deposition can be carried out in ambient temperature with no need of vacuum environment

• It is compatible to flexible substrates hence roll to roll printing on larger surfaces is possible. • It is cheap comparatively and environmental friendly.

• Functionality of organic molecules can be tailored by synthesis and adaptive tunable devices can be fabricated.

The main draw backs of organic semiconductor are its low speed and lack of long term stability. Organic light emitting diodes (OLED) is quite mature now and SONY has introduced 11´´ TV with OLED display having high contrast and almost 180◦ viewing angle.

2.1.1 Conjugated Polymers

A polymer is composed of simple repeating structural units called monomers and a conjugated

polymer is the one with alternate single and double bonds. Covalent bonds among carbon atoms define the stem (core) of the polymer and the secondary bonding between hydrogen bonds holds groups of polymers together. The chemical reaction in which monomers link together to form polymers is called polymerization. The polymerized structure can be linear, branched or ring shaped. Copolymers are formed when there is more than one type of monomers.

2.1.1.1 Molecular Structure

The main element of a polymer is carbon having electronic configuration of 1S2,2S2,2P2 , so the valence electron configuration is 2S2,2Px1,2Py1.It means that carbon atom can make only two bonds, using its two electrons in P orbital. By promotion, one electron from 2S is excited to 2Pz changing the configuration to 2S1,2Px1,2Py1,2Pz1 enabling it to make four bonds through these hybridized orbitals. The extra energy spent on promotion pays back in the shape of greater molecular stability due to two extra bonds.

Figure 2 : Formation of SP3 _{Hybridized energy levels} [2]

The SP3 hybridized levels have energy slightly larger than 2S but smaller than 2P orbital

Looking into the example of methane (CH4), experimental evidence showed that the bond angle was

109.50 _{with an overall shape of tetrahedron.}

Figure 3: Tetrahedron structure of CH4[2]

In polyethylene, the repeat unit – (CH2-CH2) n— are SP3 hybridized and each carbon atom is attached

to four other atoms through sigma (

σ

)bonds and the shared electrons lie between the overlap regions of the interacting nuclei.

Figure 4 Molecular structure of polyethylene [21]

Figure 5 Molecular structure of Trans polyacetylene [21]

Every C atom forms three sigma (

σ

) bonds with three other atoms using SP2 _{hybridization while the}

remaining one electron in the PZ orbital (perpendicular to the plane) overlap and forms a delocalized

pi(

π)

orbital.

Figure 6 : Formation of SP2 _{hybridized orbital} [2]

There exists a classification of polymers on the basis of their molecular structure. If the energy of the ground state does not change by interchanging the single and double bond in the structure, then they are called as polymers with degenerate ground state i.e. transpolyacetylene.

But if the ground state energy changes by interchanging the single and double bond character then they are called as polymers with non-degenerate ground state i.e. cis-polyacetylene

2.1.1.2 Band Structure

As is shown in Figure 8 , the sigma (

σ

) bonds lay planer defining the skeleton of the molecule, the pi (

π)

orbitals are perpendicular to the plane, providing a path for the flow of electrons.

Figure 8 Formation of sigma (

σ

) and pi (

π)

bonds [2]

The top of the filled valence band is called Highest Occupied Molecular Orbital (HOMO) and the bottom of the conduction band is called Least Un-occupied Molecular Orbital (LUMO).

Figure 9: Band gap variation as a function of length of polymer [4]

If two of the 2P atomic orbitals overlap as in Figure 9, then both of the electrons will stay in bonding molecular orbital

π

(HOMO) and the first transition level will be the anti bonding molecular orbital

π

*(LUMO).As the number of carbon atoms increase in the molecular chain, there will be N 2P molecular orbitals. These molecular orbitals are close enough and can be represented as two bands with a small band gap. In infinitely long molecule, we will have a one dimensional metal with no band gap, but Peirel explained the geometrical modification for stability removing the degeneracy of

HOMO and LUMO, resulting in a semi conducting state with a certain band gap i.e. 1.5eV for transpolyacetylene [18].

2.1.2 Doping of a Polymer

Undoped polymers have a very small value of conductivity i.e. less than 10-5S/cm for transpolyactylene. But the conductivity can be increased manifolds by doping.

The organic polymers have occupied

π

valence bandand un-occupied

π

* conduction band with a small band gap, a low ionization potential (normally less than 6eV) and a high value of electron affinity (about 2eV) as shown in Figure 9.

These are the reasons that polymers can be easily oxidized by electron accepters molecules i.e. I2,AsF5

etc and can easily be reduced by electron donors like Na, K etc. As the redox reaction is an electron transfer reaction so reduction of one molecule is accompanied by oxidation of another. So n-doping will mean reduction (addition of electron) of polymer and p-doping means oxidation (removal of electron) of polymer. In electrochemical doping, the doping charge comes from the electrodes and the doping level is determined by the applied voltage at the electrodes.

In chemical doping, the doping charge on polymer comes from another chemical , which becomes the counter ion.

Charge injection at the metal-semiconductor polymer interface (M-SC interface) can also be a method used for doping a polymer. The polymer at the interface is oxidized, when holes are injected into HOMO and reduced if electrons are injected into LUMO, resulting in an increase in the overall conductivity. But this enhanced conductivity needs constant injection (biased voltage) and polymer at the interface returns to its former state (different from electrochemical doping, where doping level remains constant) when supply voltage is turned OFF.

2.2 Charge Carriers and Charge Transport in a polymer

The Charge carriers which are produced by oxidation or reduction travel through the

π-

bonds along the polymer chains. In a conjugated polymer with non-degenerate ground state, a positive charge is created upon oxidation which resides on the polymer backbone. The charge is localized on a section of the polymer accompanied by a change from the low energy aromatic structure to the high energy quinoid structure. With another transition, an electron becomes a free radical residing between the boundaries of quinoid and aromatic energy levels. The interaction between the charge, through the distorted bond structure with the radical ion, forms a polaron (P+) having half spin and single charge.

(c) Difference of energies in Aromatic and Quinoid structures Figure 10: Charge carriers in polymers [4]

Polaron can be positive (P+) or negative (P-) depending on the oxidation or reduction of the polymer. Due to energy difference between quinoid and aromatic structure, these polarons have to overcome the barrier between the two structures to move. If the polymer is further oxidized and the existing polymer loses its electron, then the interaction between two polarons via geometric distortion can be called bipolaron (BP++) having double charge and zero spin. Bipolaron can also be positive (BP++) or negative (BP--) depending on oxidation or reduction of the polymer chain.

Formations of bipolarons are more favorable than isolated polarons beacause bipolarons decrease the ionization energy of the polymer chain.

In case of polymers with degenerate ground state like transpolyacetylene, positive charges are created which are not bound to each other due to equal energy of both bonding configurations. The two different bonding configurations are separated by domain walls which are delocalized on a few repeat units. These defects having different bonding configuration are called solitons, which can be positive (S+) or negative (S-) depending on doping.

The charge carrier on a specific chain is transported by altering the single and double bonds along the skeleton of the polymer but the material will have low conductivity as the chain length is limited. So the charge carriers move with phonon assisted tunneling by variable range hopping among different polymer chains.

The conductivity

σ

is;

σ = (n) (e) (µ)

………(1)

Where

n

is charge carrier concentration,

e

is charge on an electron and

µ

is the mobility of charge carrier.

Charge carrier concentration can be increased by heavily doping the polymer but beyond certain level, the carriers start interacting with each other hampering mobility.

Mobility is affected by many factors. Molecular order and anisotropy of the film. If the carrier

residency time is small, then molecules can not geometrically relax and trap charges and carriers move in band motion but at high temperature interacting units lose coherence and charges move by hopping decreasing the full effective bandwidth.

2.3 PEDOT: PSS as a Semiconductor

The conducting polymer poly (3, 4-ethylenedioxithiophene): poly (styrene sulfonate) or PEDOT: PSS is the most widely used material in flexible and printing electronics. The conducting PEDOT can be polymerized in PSS solution to give PEDOT: PSS. PEDOT is positively doped and the PSS acts as counter ion to balance the doping charges. It has all the qualities of being conducting, transparent and flexible. The morphology is such that conductive PEDOT: PSS resides in grains with insulating PSS making the insulating grain boundaries. Its conductivity can be increased manifolds by adding diethylene glycol (DEG) which changes its morphology. PEDOT: PSS has already taken the place of Indium tin oxide (ITO) in light emitting diode as a transparent electrode[10]_{. It is available as a thin}

film on polyester substrate called Orgacon EL-300Ω/square manufactured at AGFA.

Figure 11: Structure of PEDOT: PSS [1 2]

Doped PEDOT can have conductivity in the range of 1 to 100 S/cm and with PSS; its conductivity is around 100 S/cm [11]. As it is a low band gap material, it absorbs light in visible spectrum and converts to deep blue in color. But when it is doped, then it absorbs longer wavelengths.

2.4 Electrolytes

Electrolyte is the ionic conduction medium between the electrodes and can be in the form of liquid, solid or gelled state[13]. Examples of liquid electrolytes are water and acetonitrile. The later has broader electrochemical spectrum and can bear higher potential before decomposition.

Liquid electrolytes are used for electrochemical measurement but gelled electrolytes are preferred for devices as they are non volatile.

In liquid electrolytes, ion motion is not so swift because every ion is surrounded by solvent molecules and the whole bulk limits ionic mobility. In polymer electrolytes, ions can move through the flexible chains of the polymers suffering much less resistance [13]. Crystalline solids have strong packing and wouldn’t let the ions to move freely.

PSS based electrolyte conductivity is about 10 µS/cm while HEC based electrolyte has a conductivity of about 100 µS/cm when the relative humidity of the surrounding is 40%.

3: Basic Structures

It is important to understand the reversible redox reaction occuring inside the transistor responsible for modulating channel resistance.

The part of redox reaction from right to left is the part that is happening at the gate terminal, so a positive gate voltage is applied for this purpose, while the reaction from left to right is part of the reaction occurring at the drain terminal along the channel, requiring a negative voltage at the drain terminal.

PEDOT+PSS- + M+ + e- reduction PEDOT0 + M+ PSS- …… (2) oxidation

PEDOT+ _{is conducting and transparent in color while PEDOT}0 _{is semi-conducting with deep blue}

color.

3.1 Structure1

This structure has got a PEDOT: PSS sheet covered with an electrolyte and is also known as the Dynamic configuration or structure 1 because only one PEDOT: PSS electrode is used. When voltage is applied at the two ends of the electrode, reduction reaction takes place on the negatively biased electrode while oxidation occurs at positively biased electrode. The redox reaction occurs at the interface of PEDOT:PSS sheet and the electrolyte and hence the current has two paths to flow i.e. one is the ionic motion through the electrolyte and the other is electronic current through the PEDOT sheet. When redox reaction reaches equilibrium, then the current flows due to electrons flow only. Due to heavy concentration of undoped PEDOT, a blue front is visible at the electrode end with the negative supply. This blue front disappears as soon as the supply voltage is turned OFF.

3.2 Structure 2

Structure 2 has got two PEDOT: PSS electrodes connected through an electrolyte. When an external voltage is applied, more carriers are generated at the electrode having positive voltage and less on the negative electrode. As the two electrodes have no direct link but only through the electrolyte, only ionic current will flow as a result of the redox reactions taking place at both electrodes with the

electrolyte. Due to large concentration of undoped PEDOT, the negative electrode turns blue while the doped electrode at the positive supply remains transparent, a state that this structure retains even if the supply is turned OFF.

Figure 13: Structure 2 using two PEDOT:PSS electrodes [6]

3.3 Lateral Electrochemical Transistors (LECT)

Both structure 1 and 2 are combined together to make a novel three terminal structure of an

electrochemical transistor (ECT). Source and drain are connected to each other by structure 1 while gate terminal is connected to the channel via structure 2 using an electrolyte. As the gate lies adjacent to the channel, it is called Lateral Electrochemical Transistors (LECT). When a positive voltage is applied at the gate terminal, the PEDOT under the electrolyte is oxidized while the PEDOT in the channel under the electrolyte is reduced resulting in variation (increase) in the channel resistance. The voltage applied between gate and source controls the current through the channel.

3.4 Vertical Electrochemical Transistors (VECT)

The gate terminal can also be put up side down on the channel with electrolyte in between creating an efficient structure of Vertical Electrochemical Transistor (VECT). In vertical design, ionic motion through the electrolyte is a lot faster and the electrolyte lasts longer due to less evaporation. So VECTs can be used at a higher frequency with a longer lifetime compared to LECTs.

Figure 15: Structure of Vertical Electrochemical Transistor (VECT) [5]

4: Amplifiers 4.1 SPICE model of Electrochemical Transistor

As the VI characteristic curves of an ECT resemble that of a traditional Silicon based Depletion mode Field Effect Transistor (FET), so its terminal were also given names as Drain, Source & Gate.

Lilliehök [7]developed an equivalent RC circuitry for structure 2.

(a) Structure 2 (b) equivalent circuit for structure 2

Figure 16: Structure 2 and its equivalent circuit representation [6]

This extra RC portion for different ON/OFF time was connected to a PMOS Depletion MOSFET and the resulting circuit looked as follows with the corresponding symbol (dot showing source)

(a) Extra circuitry with a depletion PMOS (b) Symbol

Figure 17: Model of an ECT with corresponding symbol

For a positive signal the diode will turn ON and the input signal will have the parallel combination of 200k and 900k and a capacitor of 0.5µ so the time constant will be around 0.1 s. For a negative (it is also positive due to the off-set voltage) voltage, the diode will be OFF and due to the large resistance of 900k and capacitor of 0.5µ , the time constant will be about 0.45 s. As the capacitor charges quickly in the first case, the device turns OFF quickly compared to the time it takes for turning ON. As Silicon MOSFET is not symmetrical so a simple solution will be to combine two of them as follows:

(a)Symmetrical structure (b) Symbol

Figure 18: Symmetrical model of an ECT [7]

The electrochemical transistors which we used were symmetrical in the sense that the drain and source were not specified and were defined by the supply voltage.

4.1.1 Forward Characteristics Curves (FCC) & Transfer Characteristics Curves (TCC) The Forward Characteristics Curve (FCC) represents the trace of drain current (Ids) as a function of voltage between drain and source (Vds) at a specific value of gate voltage (Vgs).When no Vgs is applied, at low value of Vds, the drain current(Ids) increases linearly due to uniform impedance of the channel. But as Vds is further increased, excessive reduction reaction at the drain terminal decreases the charge carrier concentration and pinch-off occurs, hence any more increase in Vds is not

accompanied by an increase in Ids, but instead it remains constant. This highest value of the drain current is called saturation level drain to source current (Idss). Similarly, if a small Vgs is applied, the trend remains the same but the pinch-off occurs at lower value of Ids.

The Transfer Characteristics Curve (TCC) is a plot of drain current (Ids) as a function of positive gate voltage (Vgs) at a fixed value of drain to source voltage (Vds).By increasing the Vgs in steps while keeping Vds constant, the channel is more and more reduced modulating the impedance resulting in a decrease in Ids.The channel is completely blocked when Vgs=│Vp│, where Vp stands for pinch-off voltage. From the TCC, Ids and Vgs can be related to each other with Shockley’s equation:

Id= Idss [1-(Vgs/Vp)]2 ……….. (3)[8]

As Idss and Vp are constants, so the control variable is only Vgs which controls the flow of current through the channel.

(a) Vds= 0 to -3 V, Vgs= 0 to 1.2 V(0.3V step) (b) Vds= -3V(const) ,Vgs= 0 to 1.2 V(0.1V step) Figure 19: FCC & TCC of ECT model in SPICE

4.1.2 Transconductance:

The gate-to-source voltage (Vgs) controls the drain-to-source current (Ids) in an electrochemical transistor. So a change in Vgs will cause a change in the value of Ids.

∆Ids=gm ∆Vgs ……….. (4) [8]

The constant of proportionality is called Transconductance (gm).

Trans means a relation between input and output while conductance is a ratio of current-to-voltage.

So Transconductance is a ratio of change in output current (∆Ids) to a change in input voltage (∆Vgs).

gm = ∆Ids/∆Vgs ………(5) [8]

Its unit is siemen (S).

Transconductance can be calculated by slope of the transfer characteristic curve (TCC). Taking derivative of eq (2) w.r.t. Vgs, we can write;

gm = 2(Idss/Vp)[1-(Vgs/Vp)] …(6) [8]

The slope of TCC will be maximum at Vgs=0 V, so transconductance at this value will be;

gm0 = 2(Idss/Vp)……….(7) [8]

So, we can write eq (5) as;

gm = gm0 [1-(Vgs/Vp)] …….… (8) [8]

4.1.3 Input and Output Impedances

The input terminal (gate) of an electrochemical transistor is in contact with the channel through electrolyte only and hence the input impedance of an electrochemical transistor (ECT) is large due to flow of ionic current only. It means that the input impedance (Zin) will be relatively large to the source resistance and all the voltage drop will be across the gate terminal.

The output impedance (Zo) of the ECT can be found as the ratio of the change in drain-to-source voltage to the corresponding change in the drain current when ECT is in saturation.

The output impedance should be as small as possible as the drain-to-source voltage (Vds) will act as source for the load resistor and Zo will act as the internal resistance of this source.

Figure 20: FCC for screen printed VECT with MS-L for Zo calculation

In Figure 20 the Forward Characteristics Curves (FCC) are shown. We will calculate the value of Zo along three curves to see the variation in output impedance.

Vgs(V) _{Vds=1,3 V} Ids(A) at _{Vds=1,0 V} Ids(A) at ∆Vds Zo(Ω)=∆Vds/∆Ids

0V 1,45E-04 1,42E-04 3,00E-01 1,00E+05 0,3V 7,34E-05 7,19E-05 3,00E-01 2,00E+05 0,6V 2,43E-05 2,35E-05 3,00E-01 3,75E+05 Table 1: Values of Zo along three different curves of VECT

Table 1 shows that the output impedance increases as Vgs is increased

4.1.4 AC equivalent circuit

In the AC equivalent circuit of an ECT amplifier, all the voltage sources are grounded as they are at AC ground and the channel(from drain to source) is replaced with a current source gmVgs as the output

Zo = rd||Rd... (10)[8]

If rd (channel resistance) ≥ 10 Rd (series resistance with the drain)

So Zo = Rd …………. (11) [8]

For the output loop, the output voltage Vo will be;

Vo = - gmVgs [rd||Rd]

But Vgs = Vi (input voltage) So voltage gain (Av = Vo/Vi) will be;

Av = - gm [rd||Rd] ……… (12) [8]

In all our circuits we called Rd as Rs (series resistance) and gm [rd||Rs] will be our calculated voltage

gain value.

4.2 Various Amplifier circuits

The following three amplifier circuits were used primarily to test ECT as an amplifier. The common source configuration is shown in Figure 22.

It is the simplest circuit for an amplifier having only one resistance at the drain. An input signal is applied at the gate with some off-set voltage and output is measured at the drain of the ECT.

Figure 22: simple amplifier

V1 & V2 are used for the measurement of input and output voltage measurement. The input voltage applied at the gate terminal (Vin=Vgs) varies the resistance in the channel and hence the voltage drop from drain to source (Vds) varies accordingly.

For the output loop;

-VS= -IcRs – Vds ………. (13) [8]

VS= applied voltage

Ic= current through the channel Rs= Series resistance

As Vgs rises, the channel resistance increases, so current Ic decreases and hence Vds will become more negative. Hence the input Vgs and output Vds are out of phase from each other.

Figure 23 also shows a common source configuration, but now the source terminal is not at dc ground due to the presence of a resistor but at ac ground due to the presence of the capacitor.

VM1 & VM2 are used for the measurement of input and output voltage measurement.

The introduction of an extra resistance at the source terminal helps in pushing the operating point of the transistor towards the middle making it possible to get a voltage gain (A) equal to the ratio of the two resistors, if there is no capacitor.

A= R1/R2 ………….. (14) [8] VIN 0 R1 IVm2 IVm1 R2 V1 3 1m _C1

Figure 23: An amplifier circuit having a bypass capacitor

The capacitor bypasses the ac signal from the flowing through the 2nd resistance R2, keeping the source at ac ground. Capacitor can very easily be manufactured by putting two PEDOT sheets on top of one another with electrolyte in between them.

The circuit shown in Figure 24 has a current source in the output loop. As the gate and source terminals are at the same voltage in the ECT in the out loop, it will have a constant current flowing through it.

When an input voltage (Vin) is applied at the input of the other ECT, it modulates channel resistance affecting the voltage V2.

This variation will not affect the current through the ECT in the output loop because this small

change in Vds does not affect the current as it is operating in saturation region. So if the channel resistance has increased and the current remains the same, then we will get a large swing in voltage V2 according to Ohm’s law.

By careful selection of the two resistances, we can achieve a high gain with small values of R1 & R2 and Vos (offset voltage).

A differential amplifier is shown in Figure 25. Two ECTs are used with their drains connected to the supply via resistors R1 and R2, where R1=R2.

The input signal is applied at the gate of one of the ECT and this difference signal between the two gate terminals is amplified which is measured at the drain terminals by using a voltmeter Vm1.

ECT1 ECT2 R3 V0 3 R1 V1 IVm1 R2 V2 +805.17m IVm2

Figure 25: Differential amplifier

For a differential amplifier, we normally need a matched transistor pair having equal current gain, so that with R1=R2 on the drain terminals, we will have both the terminals at the same voltage if the same input is applied at both terminals.

So the difference of voltage measured through Vm1 will be zero. Hence for the same input signals, we will have zero output meaning that common signal is not amplified at all.

But if different signals are applied at both inputs, then their respective drain terminals will be at different voltages and Vm1 will measure a large output voltage. Noise signals can be completely eliminated because that will appear as the same input to both transistors.

The problem with electrochemical transistors is to find a matched pair which is almost impossible. So what we can do is to adjust the off-set voltages such that the output is at zero level meaning same voltages at both the drain terminals. An amplified replica of input voltage signal applied after this can be predicted by Vm1. Hence only difference between the input signals is amplified using the

differential amplifier.

5: Experimental Part 5.1 Dimensions of ECTs

While designing the EC Transistor, the following were the various parameters: Effective channel area having electrolyte=Ac=300 µm*300 µm

Effective gate area having electrolyte=Ag Gate area (Ag) ≥5*Effective channel area (Ac)

Drain & Source height=H= 5mm Drain & Source length=L= 5mm

Figure 26: Dimensions of ECT made from plotter technique

5.2 Manufacturing Techniques 5.2.1 Plotter technique

• Make the layout design (transistor+lamination) in Corel Draw 9. • Put the PEDOT sheet on the plotter and adjusted the two areas. • By adjusting the force of the cutter, executed the cutting the PEDOT.

• Took the lamination foil and put adhesive tape on both ends and patterned it by plotter. • Took the two layers of lamination and put it on PEDOT and passed it through the laminator. • Put carbon contacts on Drain, Gate and Source.

• Baked the sample at 80o _{C for 10 minutes.}

• Applied the electrolyte and removed the top layer of the lamination sheet leaving only the lamination foil on PEDOT.

• Baked the sample for about 2 minutes at 60o_{C, so that the electrolyte forms an even layer.}

• Put the sample in ambient temperature to stabilize and adapt to the humidity of the surroundings.

• For Vertical design as shown in Figure 15, the gate terminal is put on top of the channel with electrolyte in between. Lamination or adhesive tape is needed to keep the gate and channel in contact.

5.2.2 Screen Printing Technique:

Screen printing technique is a commercial method of roll to roll printing. The viscous ink is put on top of thin mesh of wires and squeegee pushes the ink through the wires onto the substrate lying

underneath. Resolution is up to 75µm while 30*30 cm sheet is printed in 5seconds [9].

Figure 27: Screen printing technique [9]

The manufacturing steps of screen printed electrochemical transistors were: • PEDOT layer was printed on plastic substrate

• Carbon was applied for contacts

• A lacquer layer was printed on top of PEDOT layer to define the area for electrolyte • Electrolyte was applied as a final step in the clearance between the carbon

electrodes.

5.3 Experimental Setup & Data Analysis

After manufacturing, the output (Ids vs. Vds at certain levels of Vgs) and transfer (Ids vs. Vgs at fixed value of Vds) characteristic curves were checked with Parameter Analyzer. For output characteristic curves, Vds was varied from 0 to -3V while Vgs was varied from 0 to 1.2V (with a step of 0.3V). For transfer characteristic curves, Vgs was varied from 0 to 1.2V with Vds=-2V.

After verifying the transistors, variable resistors were mounted on a circuit board and ECT was connected through clips. Input signal at a fixed frequency and with certain offset was applied from a function generator, supply voltage was given from a DC power supply and input/output waveforms were studied on an oscilloscope. In parallel to this, data was also saved using LABVIEW program on a PC connected to the circuit via data acquisition card.

The recorded data was plotted using MS Excel.

5.3.1 Time Response Measurement:

For the time response measurement, a square wave voltage was applied with a peak value of 1V at the gate and the corresponding changes in drain-to-source current measured. The drain-to-source voltage was kept constant at -1V.The gate voltage was applied in 0-1-0 with every state maintained for 2.5 seconds and the same sequence was repeated ten times with a time delay of about 0.5 seconds.

Input at Gate(Vgs) 0,0 0,2 0,4 0,6 0,8 1,0 0 4 8 12 16 Time(s) Vgs( V) Vgs(V)

6: Results and Discussions 6.1 Lateral Electrochemical Transistors (LECT):

6.1.1 LECT with EG010

FCC for LECT with EG010(Sample4,Lot1)

-1,2E-04 -1,0E-04 -8,0E-05 -6,0E-05 -4,0E-05 -2,0E-05 0,0E+00

-3,0E+00 -2,5E+00 -2,0E+00 -1,5E+00 -1,0E+00 -5,0E-01 0,0E+00 Vds(V)

Id

s(

A

)

TCC for LECT with EG010(Sample 4,Lot1)

-1,2E-04 -1,0E-04 -8,0E-05 -6,0E-05 -4,0E-05 -2,0E-05 0,0E+00

0,0E+00 2,0E-01 4,0E-01 6,0E-01 8,0E-01 1,0E+00 1,2E+00

Vgs(V)

Id

s(

A

)

(a)Vds =0 to -3V & Vgs= 0 to 1.2 V (b) Vgs= 0 to 1.2V & Vds= -3V

FCC of LECT with EG010(Sample 4 ,Lot2)

-1,6E-04 -1,2E-04 -8,0E-05 -4,0E-05 0,0E+00

-3,0E+00 -2,5E+00 -2,0E+00 -1,5E+00 -1,0E+00 -5,0E-01 0,0E+00

Vds(V)

Id

s

(A

)

TCC for LECT with EG010(Sample 4,Lot2)

-2,5E-04 -2,0E-04 -1,5E-04 -1,0E-04 -5,0E-05 0,0E+00

0,0E+00 2,0E-01 4,0E-01 6,0E-01 8,0E-01 1,0E+00 1,2E+00

Vgs(V) Id s (A ) (c)Vds =0 to -3V & Vgs= 0 to 1.2 V (d) Vgs= 0 to 1.2V & Vds= -3V Figure 29: (a) & (b) are for one sample while (c) & (d) are for another one.

The Forward Characteristics Curves (FCC) and Transfer Characteristics Curve (TCC) for two different LECTs are shown in Figure 29. A delay of 3 seconds was introduced between consecutive readings to get a the device in a stable condition before taking reading.

Time Response for LECT with EG010 -1,50E+00 -1,00E+00 -5,00E-01 0,00E+00 5,00E-01 1,00E+00 0 20 40 60 80 100 Time(s) V o lt ag e( V ) -7,0E-05 -3,5E-05 0,0E+00 3,5E-05 7,0E-05 Id s( A ) Vgs

Ids Time Response for LECT with EG010

-1,5E+00 -1,0E+00 -5,0E-01 0,0E+00 5,0E-01 1,0E+00 0 20 40 60 80 100 Time(s) Vo lt ag e( V) -1,0E-04 -5,0E-05 0,0E+00 5,0E-05 1,0E-04 Id s( A ) Vgs(V) Ids(A)

(a)Vgs= 0 & 1. 5V while Vds= -1V (b)Vgs= 0 & 1. 5V while Vds= -1V Figure 30: Time response of the two LECTs with EG010 as electrolyte

The Time response for an LECT with EG010 is shown in Figure 30 . As the Vgs (square waveform) was applied keeping Vds at a fixed voltage, the ON current through the device decreases at every cycle. As the blue front develops with time, the device takes more time to turn ON and remains OFF (current remains zero) for most of the time.

When LECT with EG010 as electrolyte were checked in the simple amplifier circuit of Figure 22, it showed a small gain of 2.5 and 3.

LECT5 Self made (A=2.5)

-2 -1,5 -1 -0,5 0 0,5 1 0 50 100 150 200 250 300 Time(s) V o lt ag e( v) Input output

LECT sample4 with EG010(A=3)

-2 -1,5 -1 -0,5 0 0,5 1 0 50 100 150 200 250 300 Time(s) Vo lt ag e( V) Input Output (a)VS=3V,Vin=200mVpk-pk,Vos=620mV,f=50mHz,R=64k (b)VS=3V,Vin=50mVpk-pk,f=30mHz,Vos=700mV,R=382K

Figure 31: LECT with EG010. Both of the samples represented in this diagram had EG010 as their electrolyte.

The decrease in the operating voltage for LECT5 shows an increase in the drain resistance, which is due to the blue front effect becoming prominent at every cycle until reaching a maximum level.

6.1.2 LECT with MS-L

Four samples of LECTs were prepared by plotter technique and MS-L was used as electrolyte.

FCC for LECT with MS-L

-1,6E-04 -1,2E-04 -8,0E-05 -4,0E-05 0,0E+00

-3,0E+00 -2,4E+00 -1,7E+00 -1,1E+00 -4,1E-01

Vds(V) Id s( A ) sample1

TCC for LECT with MS-L

-2,0E-04 -1,5E-04 -1,0E-04 -5,0E-05 0,0E+00

0,0E+00 3,0E-01 6,0E-01 9,0E-01 1,2E+00

Vgs(V) Id s( A ) sample1

(a)Vds =0 to -3V & Vgs= 0 to 1.2 V (b) Vgs= 0 to 1.2V & Vds= -3V FCC for LECT with MS-L

-2,5E-04 -2,0E-04 -1,5E-04 -1,0E-04 -5,0E-05 0,0E+00

-3,0E+00 -2,0E+00 -1,0E+00 0,0E+00

Vds(V) Id s( A ) sample2

TCC for LECT(sample2 with MS-L)

-3,0E-04 -2,3E-04 -1,5E-04 -7,5E-05 0,0E+00

0,0E+00 4,0E-01 8,0E-01 1,2E+00

Vgs(V) Id s( A ) sample2 (c)Vds =0 to -3V & Vgs= 0 to 1.2 V (d) Vgs= 0 to 1.2V & Vds= -3V Figure 32: (a) & (b) are for one ECT while (c) & (d) are for another sample.

From the FCC and TCC for first transistor, it is evident that highest value of drain current (Idss) is higher in TCC. The reason can be the fact that in TCC it was in the first reading which measures Idss and blue front has not yet been formed.

Time Response of LECT (sample1 with MS-L) -1,0 -0,5 0,0 0,5 1,0 Vgs (V) -6,0E-05 0,0E+00 6,0E-05 1,2E-04 Id s( A ) Vgs(V)

Ids(A) Time Response of LECT

(sample2 with MS-L) -1,0 -0,5 0,0 0,5 1,0 Vg s( V) -1,0E-04 0,0E+00 1,0E-04 2,0E-04 Id s( A ) Vgs(V) Ids(A)

The time response was quite promising with MS-L as electrolyte and the channel could be turned ON and OFF completely by the applied Vgs.

LECT(sample1) with MS-L(A=10)

-1,5 -1 -0,5 0 0,5 1 1,5 0 50 100 _Time(s)150 200 250 300 V o lt ag e( V ) Input Output

LECT(sampl2) with MS-L(A=8.7)

-2 -1,5 -1 -0,5 0 0,5 1 0 200 400 600 Time(s) V o lt ag e( V ) Input Output

(a)VS=3V,Vin=50mVpk-pk,Vos=1V,f=30mHz,Rs=414k (b)VS=3V,Vin=50mVpk-pk,Vos=0.8V,f=30mHz,Rs=190k

LECT(sample1) with MS-L(A=8.8)

-2 -1,5 -1 -0,5 0 0,5 1 1,5 0 100 200 300 Time(s) V o lt a g e( v) Input Output LECT(sample2) with MS-L(8.7) -2 -1,5 -1 -0,5 0 0,5 1 1,5 0 50 100 150 200 250 300 Time(s) Vo lt a g e (V) Input Output (c)VS=3V,Vin=50mVpk-pk,Vos=1V,f=30mHz,Rs=192k (d)VS=3V,Vin=50mVpk-pk,Vos=1.1V,f=30mHz,Rs=257k

Figure 34: LECT with MS-L. (a) and (b) are tests on the same day of manufacturing while (c) and (d) tests on the 7th day of manufacturing

Two of the transistors were used in the simple amplifier circuit of Figure 22 and the results were quite promising giving us a gain of 10 and 8.7.When the same circuits were checked one week later, we could get the almost the same gain but at other values of series resistance (sample 1 & 2) and offset voltage (sample 2), which gives a hint about the change in the operating point as the electrolyte dries.

6.2: Vertical Electrochemical Transistor (VECT) 6.2.1 VECT with EG010

Four VECT samples were made with plotter technique using EG010 as electrolyte. FCC for VECT with EG010

-1,6E-04 -1,2E-04 -8,0E-05 -4,0E-05 0,0E+00

-3,0E+00 -2,3E+00 -1,5E+00 -7,5E-01 0,0E+00

Vds(V)

Id

s(

A

)

(a)Vds =0 to -3V & Vgs= 0 to 1.2 V (b) Vgs= 0 to 1.2V & Vds= -3V FCC for VECT with EG010

-1,2E-04 -9,0E-05 -6,0E-05 -3,0E-05 0,0E+00

-3,0E+00 -2,3E+00 -1,5E+00 -7,5E-01 0,0E+00 Vds(V) Id s( A ) (c)Vds =0 to -3V & Vgs= 0 to 1.2 V (d) Vgs= 0 to 1.2V & Vds= -3V Figure 35: FCC & TCC for two VECTs with EG010 as electrolyte

TCC and FCC have different values of Idss due to the reason that in FCC, this is the first reading while in FCC there is a blue front developed before reaching to that measurement.

Time Response of VECT with EG010

-1,0E+00 -5,0E-01 0,0E+00 5,0E-01 1,0E+00 Vg s (V ) -5,0E-05 0,0E+00 5,0E-05 1,0E-04 1,5E-04 Ids (A ) Vgs(V)

Ids(A) Time Response of VECT with EG010

-1,0 -0,5 0,0 0,5 1,0 Vgs( V) -4,0E-05 0,0E+00 4,0E-05 8,0E-05 Id s( A ) Vgs(V) Ids(A)

The time response shows that there is a blue front developing as the device is ON for more time,which reduces the ON current significantly.

These VECTs were used in simple amplifier circuit of Figure 22.Two of the results are shown the following figure indicating considerable gains of 6 and 8.54.

Self made VECT,sample1(A=6)

-1,5 -1 -0,5 0 0,5 1 0 50 100 150 200 250 300 Time(s) V o lt ag e( V ) Input Output

Self made VECT,sample2(A=8.54)

-2 -1,5 -1 -0,5 0 0,5 1 0 50 100 150 200 250 300 Time(s) V o lt ag e( V ) Input Output

(a)Vs=3V,Vin=50mVpk-pk,Vos=0.65V,f=30mHz, Rs=142k (b)VS=3V, Vin=50mVpk-pk, Vos=0.76V, f=30mHz, Rs=142k

Figure 37: Input/output curves for VECTs used in amplifier circuit

Both of the results shown here indicate that VECTs with EG010 can also be used for amplification but at much lower frequencies.

6.2.2 VECT with MS-HEC

Four VECT samples with MS-HEC were made. They were checked in the simple circuit of Figure 22.

Sample2 with MS-HEC A=7.96(t=100s),A=7.8(t=300s) -1,5 -1 -0,5 0 0,5 0 50 100 150 200 250 300 350 400 Time(s) V o lt ag e( V ) Input Output

Sample4(run2) with MS-HEC A=7.44(t=100s) -1,5 -1 -0,5 0 0,5 1 0 50 100 150 200 250 Time(s) Vo lt ag e( V) Input Output (a)VS=3V,Vin=50mVpk-pk,Vos=300mV,f=30mHz,Rs=300k (c)VS=3V,Vin=50mVpk-pk,Vos=750mV,f=30mHz,Rs=307k

Sample2 with MS-HEC(one day later) A=5(t=60s) -2 -1,5 -1 -0,5 0 0,5 1 0 50 100 150 200 250 300 Time(s) Vo lt ag e( V) Input Output

Sample4 with MS-HEC(one day later) A=0.7(t=100s) -0,5 -0,2 0,1 0,4 0,7 1 0 50 100 150 200 Time(s) Vo lt ag e( V) Input Output (b)VS=3V,Vin=50mVpk-pk,Vos=400mV,f=30mHz,Rs=400k (d)VS=3V,Vin=50mVpk-pk,Vos=750mV,f=30mHz,Rs=307k

Figure 38: Self made VECT with MS-HEC

Part (a) and (c) are the results of the same day when the VECT were made and the results were quite encouraging. But only after one day the results deteriorated a lot under the same circuit parameters. Offset (Vos) and Series resistance (Rs) values were changed but even then, the results were not satisfactory enough as shown in (b) and (d).

6.2.3 VECT with MS-L

Four VECTs were manufactured by the plotter technique with MS-L as electrolyte.

FCC for VECT sample2 with MS-L

-1,8E-04 -1,6E-04 -1,4E-04 -1,2E-04 -1,0E-04 -8,0E-05 -6,0E-05 -4,0E-05 -2,0E-05 0,0E+00

-3,0E+00 -2,5E+00 -2,0E+00 -1,5E+00 -1,0E+00 -5,0E-01 0,0E+00

Vds(V) Id s( A ) Ids

(a)Vds =0 to -3V & Vgs= 0 to 1.2 V (b) Vgs= 0 to 1.2V & Vds= -3V FCC for VECT sample4 with MS-L

-3,5E-04 -3,0E-04 -2,5E-04 -2,0E-04 -1,5E-04 -1,0E-04 -5,0E-05 0,0E+00

-3,0E+00 -2,5E+00 -2,0E+00 -1,5E+00 -1,0E+00 -5,0E-01 0,0E+00

Vds(V) Id s (A ) (c)Vds =0 to -3V & Vgs= 0 to 1.2 V (d) Vgs= 0 to 1.2V & Vds= -3V Figure 39:FCC and TCC with MS-L as electrolyte

The time response of a VECT with MS-L is shown in the following figure.

Time Response of VECT with MS-L

-1,5 -1,0 -0,5 0,0 0,5 1,0 0 15 30 45 60 75 90 105 Time(s) Vg s (V) -1,2E-04 -6,0E-05 0,0E+00 6,0E-05 1,2E-04 Id s (A ) Vgs(V)

Ids(A) Time Response of VECT with MS-L

-1,5 -1,0 -0,5 0,0 0,5 1,0 0 20 40 60 80 100 Time(s) Vg s (V ) -2,0E-04 -1,0E-04 0,0E+00 1,0E-04 2,0E-04 Ids (A ) Vgs(V) Ids(A)

(a)Vds= -1V(fixed),Vgs=0 & 1V (b)Vds= -1V(fixed),Vgs=0 & 1V Figure 40: Time response of VECTs with MS-L

The time response was quite promising and the device turned ON at almost the same current as in the first cycle meaning that the blue front was that effective and the electrolyte had better ionic movement than earlier ones.

These VECTs were checked in the simple amplifier circuit of Figure 22.The following diagram shows the plot of input/output of two samples.

VECT sample2 with MS-L A=7.7(t=100s),A=8.44(t=600s) -2 -1,5 -1 -0,5 0 0,5 1 1,5 0 100 200 300 400 500 600 700 Time(s) V o lt ag e( V ) input output

VECT Sample4 with MS-L A=7(t=600s) -1,5 -1 -0,5 0 0,5 1 1,5 0 100 200 300 400 500 600 700 Time(s) V o lt age(V ) input output (a)VS=3V,Vin=50mVpk-pk,Vos=880mV,f=30mHz,Rs=160k ©VS=3V,Vin=50mVpk-pk,Vos=880mV,f=30mHz,Rs=40k

Sample2 with MS-L after one week A=8.2 -2 -1,5 -1 -0,5 0 0,5 1 1,5 0 100 200 300 400 500 Time(s) V o lt ag e(V ) Input Output

Sample4 after one week A=7 -2 -1,5 -1 -0,5 0 0,5 1 1,5 0 50 100 150 200 250 300 Time(s) V o lt ag e( V ) Input Output (b)VS=3V,Vin=50mVpk-pk,Vos=880mV,f=30mHz,Rs=160k (d)VS=3V,Vin=50mVpk-pk,Vos=880mV,f=30mHz,Rs=65k

(a) &(c) are tests after one day while (b) & (d) are tests performed after one week of manufacturing

Figure 41: VECT with MS-L

As we can see these VECTs did not deteriorate much within one week. The simulation results were almost the same for sample 2 but differed for sample 4.

6.2.4 Frequency Response

The following figure shows the Frequency response of Vertical Electrochemical Transistors (VECT) with MS-L as electrolyte. Three VECTs were checked in simple amplifier circuit of Figure 22 and output was checked at various frequencies with other circuit parameters fixed at the following values:

• Sample 2:VS=3V,Vin=50mVpk-pk,Vos=880mV,Rs=160k • Sample 3:VS=3V,Vin=50mVpk-pk,Vos=880mV,Rs=30k • Sample 4:VS=3V,Vin=50mVpk-pk,Vos=880mV,Rs=40k

The circuit was turned OFF for five minutes after every reading as it was our previous observation that the VECT deteriorates if used for longer times. For every reading the circuit remained ON for a few input cycles.

Frequency(Hz) Gain(A) of Sample 2 Gain(A) of Sample 3 Gain(A) of Sample 4

0,01 8,64 6,2 6,5 0,03 8 6,36 6,84 0,06 7,96 6,2 6,6 0,09 7,7 6 6,6 0,15 7,3 5,7 6,5 0,3 6,64 5,52 6,5 0,5 6,5 5,42 6,4 1 6,24 4,9 6,1 2 4,8 3 5,6

Table 2: Gain of three VECT samples with MS-L checked at different frequencies Frequency Response of VECT with MS-L

0

2

4

6

8

10

0

0,5

1

1,5

2

2,5

Frequency(Hz) Ga in (A ) Sample 2 Sample 3 Sample 4

Figure 42 Frequency response of VECT samples with MS-L

The gain is decreased at higher frequencies because the movement of ions in the electrolyte is the limiting factor.

6.3 Screen Printed Vertical Electrochemical Transistors (SP VECTs) with MS-L 6.3.1 FCC and TCC

The screen printed VECTs were used to check their performance and to look into the possibility of mass production of these devices.

(a)Vds =0 to -3V & Vgs= 0 to 1.2 V (b) Vgs= 0 to 1.2V & Vds= -3V

(c)Vds =0 to -3V & Vgs= 0 to 1.2 V (d) Vgs= 0 to 1.2V & Vds= -3V Figure 43: FCC & Tcc of Screen Printed VECTs with MS-L

Although there is some difference in the value of Idss in FCC and TCC, the time response was excellent in all the transistors checked so far. The rise and fall time were almost 0,1 second and the channel turned ON at almost the same current as in the first cycle.

Amplifiers:

(a)VS=-3V,Vos=950mV,f=30mHz,Vin=50mVpk-pk,Rs=263k (b)VS=-3V,Vos=1V,f=30mHz,Vin=50mVpk-pk,Rs=215k

(c)VS=-3V,Vos=950mV,f=30mHz,Vin=50mVpk-pk,Rs=263k (d)VS=-3V,Vos=1V,f=30mHz,Vin=50mVpk-pk,Rs=215k

Figure 45: amplifier response with screen printed VECTs. (a) & (b) are tests on the same day of manufacturing while (c) & (d) are tests of the same ECTs after one week.

6.3.2 Frequency Response

Three screen printed VECTs were checked for frequency response using the following values; • sample 1,VS=-3V,Vos=950mV,f=10Hz,Vin=50mVpk-pk,Rs=263k

• sample 4 ,VS=-3V,Vos=1V,f=10Hz,Vin=50mVpk-pk,Rs=215k

• sample 5,VS=-3V,Vos=1.1V,f=10Hz,Vin=50mVpk-pk,Rs=263k

Table 3 shows the gain of the simple amplifier circuit of Figure 22, when three screen printed transistors were used.

Gain(A) Frequency(Hz)

Sample1 Sample4 Sample5 0,01 9 9,8 14 0,03 8,3 8,54 11 0,06 9,1 9,62 13,64 0,09 9,1 9,32 13,52 0,15 9,2 9,24 13,9 0,3 8,84 9,14 13,96 0,5 7,52 8,2 13,32 1 7,4 8 11,92 2 7,5 8 2 5 4,5 5,9 0,8 8 1,5 2 0,6 10 1 1,9 0,5 Table 3: Gain of three screen printed VECTs with MS-L at different frequencies

Figure 46: Plots of gain variation as a function of input frequency

6.4 Linearity Check

For a particular amplifier circuit, the part of input for which the output will be a linear multiple of the input must be known before use. In other words, it is always desirable to know the exact area of operation, for which we can get maximum symmetrical swing of the output voltage. Maximum change in the output is detected by sweeping the input voltage in specific limited values. We can define limits of the input by taking values for which the slope of transfer curve will change by 5%.

Linearity check for VECT sample 3

0 0,5 1 0 50 100 Time(s) Input V o lt ag e, V i (V ) -3 -2 -1 0 O u tput V o lt ag e, V o (V ) input

output output vs input for two cycles

-3 -2 -1 0 0 0,2 0,4 0,6 0,8 1 Vin(V) Vo( V )

(a) output & input vs. time(Vs=-3V,Rs=200k,f= 20mHz) (b) Plot of output vs. input

Output vs Input for two cycles

-1,2 -0,9 -0,6 -0,3 0 0 0,2 0,4 0,6 0,8 1 Vin(V) Vo (V )

(c) output & input vs. time(Vs=-1,5V ,Rs=100k,f= 20mHz) (d) Plot of output vs. input

Figure 47: (a) & (b) are for self made VECT with MS-L while (c) & (d) are for SP VECT with MS-L

The transistors were connected in the simple amplifier circuit of Figure 22 and a ramp input was applied at the input with a value of 1Vpk-pk , an off-set of 500mV at a frequency of 20mHz to keep the input positive all the time and setting the symmetry to 50 % to have equal time for ramp up and ramp down.

In Figure 47 part (b), it can be seen that in self made VECT the trajectories of the output voltage are different in response to the variation in the input i.e. as the input voltage increases to modulate channel resistance to turn OFF the device, a blue front develops at the drain terminal due to reduction front migrating from the channel where ionic mobility is low and this increased resistance causes the channel to turn ON slowly [16].

Due to this reason I defined the window for increasing and decreasing input.

From the values of output voltage (Vo) and input voltage (Vin), the difference of two consecutive readings of Vo was divided by corresponding difference of Vin for the whole region where output was varying quickly. A point of maximum change was selected and then one point on either side was

chosen with a fixed 5% slope variation. The window defined in this manner gives us the limits of input variation for getting good amplification.

The values of input and output voltages from Figure 47 can be seen in Table 4.

Type Vin1 Vo1 Vin2 Vo2 |∆ Vin| |∆ Vo| A=|∆ Vo|/|∆ Vin|

0,630V -1,001V 0,750V -2,014V 0,120V 1,013V 8,44(8,6 from Excel) VECT 0,530V -0,295V 0,815V -2,373V 0,285V 2,078V 7.29(8,17 from Excel) 0.862V -0,249V 0,940V -0,857V 0,078V 0.608V 7,8(7,77 from Excel) SP VECT 0,847V -0,217V 0,952V -0,901V 0,105V 0,684V 6.5(7,35 from Excel)

Table 4: The band of input for which the output has highest variation

A straight line was fit as trend line on the data set selected and slope was determined by the straight line equation. Similarly, another data set was chosen by taking more data points on either sides and again slope was determined and changed the data set slightly, until we got exactly 5% variation in slope. For self made VECT, the output of the amplifier will remain in linear region (between -295mV and -2.737V) for an input variation of 285mV (530-815mV). Similarly, for screen printed VECT, the output of the amplifier will remain in linear region (between -217mV and -901mV) for an input variation of 105mV (847-952mV).

The hysteresis in the self made VECT was further investigated for by running the device again at a lower frequency.

Linearity check for VECT sample 3 with MS-L

0 0,25 0,5 0,75 1 0 100 200 300 400 500 Time(s) In p u t V o lt ag e, V i (V ) -3 -2,25 -1,5 -0,75 0 Ou tp u t V o lt ag e, V o (V ) input

output output vs input(for two cycles)

-3 -2,5 -2 -1,5 -1 -0,5 0 0 0,2 0,4 0,6 0,8 1 Vin(V) Vo (V )

(a) output & input vs. time(Vs=-3V,Rs=200k,f= 4mHz) (b) Plot of output vs. input

Figure 48: (a) plot of ramp input and corresponding output (b) plot of output vs. input

Figure 48 (b) shows that the effect is still there as in Figure 47 (b) although the frequency has been decreased to 4mHz meaning that the blur front effect is quite prominent even at such a small frequency.

6.5 Slew Rate Measurement

The slew rate is an important characteristic of an amplifier and can be defined as the maximum rate of change of output voltage.

Slew Rate = max [d (Vo)/ dt] ………….. (14)[22] where Vo is the output voltage (Vds in our case).

In our measurements, the amplifier circuits with screen printed VECTs had the maximum slew rate of 0,056 V/s. The amplifier circuits having ECTs with MS-L as electrolyte were meeting the general requirement of equation (15) and the slew rate was constant for rising and falling edges. For other amplifier circuits with ECTs having EG010 as electrolyte, slew rate was different for rising and falling edges and only falling edges could meet the requirement of equation (15) as shown in Table 7 of appendix.

6.6 Comparisons of self made and Screen printed transistors:

The data for all the transistors were thoroughly analyzed. Although the transistors were locally manufactured and not each one was up to the standard but still we can get a general feeling of

differences in between various structures and materials. The three types of transistors can be compared on the basis of the following parameters other than gain (detailed measurements available in

appendix).

• ON/OFF ratio: It is the ratio of the maximum ON current to the minimum OFF current. General form is;

ON/OFF ratio = (Ids with Vgs = 0V)/ (Ids with Vgs = 1V) …….. (14) Where Vds is at a constant level.

A transistor having good ON/OFF ratio is best suited to be used as an amplifier.

It was fully electrolyte dependant and Lateral, Vertical and screen printed transistors with a good electrolyte like MS-L had their ON/OFF ratio around 300, but Screen printed transistors had better average value while LECT & VECT with EG010 had an average value of less than hundred.

• Rise & Fall Time: The rise time was calculated from the time response curve as the time taken by a transistor to reach from 10 to 90 % of its maximum ON current.

Similarly, fall time was measured as the time taken by a transistor to reach from 10 to 90 % of its OFF current. The best rise & fall time was on average at 0.1 s for screen printed transistors while LECT with EG010 had the worst values of above one second.

• Maximum ON/Minimum OFF current: As all transistors were checked for ten cycles, the deviation from 2nd to 10th cycle among all the transistors was analyzed for their maximum ON and minimum OFF current.

Screen Printed transistors & VECT with MS-L had maintained their maximum ON and minimum OFF current while LECT with EG010 was worst in all having huge differences in these values for all ten cycles.

• Transconductance: As discussed earlier, the ratio of the change in channel current to the corresponding change in gate voltage is called Transconductance. A transistor having higher value of transconductance behaves as an efficient amplifier.

The average value was 0.0002 siemens for almost all the transistors due to the fact that their transfer characteristics curves were plotted by giving a delay of two seconds between two consecutive measurements while testing on the parameter analyzer.

6.7 Change in response for certain variables

All electrochemical transistors are frequency dependant. The speed of these transistors is a lot slower than traditional inorganic transistors. The reasons for their slow speed are the movement of ions as charge carriers and bulk switching for depleting the channel of the carriers.

The operation of these transistors is based on electrochemical reaction causing the doping and

dedoping of the active material, so the voltage across the drain-source terminal is important to be at a certain point. If we will change the supply voltage to a smaller value while keeping all others constant, the gain will be changed in accordance with the variation in the drain-to-source voltage.