Silicone-based Carbon Black Composite for Epidermal Electrodes

(1)

Silicone-based Carbon Black Composite for Epidermal

Electrodes

Thesis for the degree of Master of Technology

Authors: Melika Eklund, Nellie Kj¨ all Supervisor: Prof. Klas Hjort

Department of Engineering Science, Division of Microsystems Technology

Subject reader: Prof. Hugo Nguyen Department of Engineering Science, Division of Microsystems Technology

December 2019

(2)

List of Figures

1.1 Origami, a) 3-D sub-microscale pop-up silicon structure assem- bled by compressive buckling. b) Graphene origami by thermore- sponsive self-folding. c) Graphene kirigami by external force and

actuation [1]. . . . 7

1.2 Tattoo-like epidermal patch [2] . . . . 8

2.1 Layers of skin [3] . . . . 11

3.1 Carbon black primary, aggregates and agglomerations structure [4] 15 3.2 C65 TEM pictures (a) CB primary particles size 40 nm in aggre- gates, (b) CB agglomerations [5]. . . . 15

4.1 Electric vacuum and mixer (HD device) . . . . 18

4.2 Mixing process . . . . 18

4.3 The mold, which used while curing the samples . . . . 19

4.4 Three devices to measure resistance (a) RCL multimeter, (b) Key-Sight multimeter, (c) Simple digital multimeter . . . . 19

5.1 Mixing by syringe, (a) syringe connection, (b) a lot of bubbles on the surface of cured sample . . . . 22

5.2 Cracked Sample of 14 wt% carbon black still in their mold . . . . 23

5.3 Tensile test for sample 8 wt% CB + 5 wt% Ag/AgCl using ruler and stretching by hand. (a) Dog-bone mold, (b) Unstretched sample, (c) Stretched sample . . . . 25

5.4 Average of measured resistivity of 3 series of samples . . . . 26

5.5 Temperature dependency of resistivity . . . . 27

5.6 Resistance measurement on top and bottom side of samples, (a) batch 1, (b) batch 2, (c) batch 3 . . . . 28

5.7 Bulk resistance measurement with Galinstan droplet and Cu- tape, (a) Schematic view of measurement setup, (b) Resistance measurement by simple digital multimeter . . . . 30

5.8 Average of measured resistivity of batches of 3 samples . . . . 31

(4)

5.9 Students’ T-test (statistical test) to check the range of possible results, the black dots are resistivity experimental values for dif- ferent samples with the same compositions, black circles are mean values and errorbars are showing the confidence intervals that is the range, where the results are probable to occur as high as 95%. 33 7.1 Reisistivity vs filler concentration,RCL multimeter, hand-cut sam-

ples . . . . 42 7.2 Reisistivity vs filler concentration,Simple digital multimeter, Laser-

cut samples . . . . 42 7.3 Reisistivity vs filler concentration,RCL multimeter and hotplate,

Laser-cut samples . . . . 43

(5)

List of Tables

3.1 Properties of elastomers used in this paper . . . . 13 3.2 Dilutant properties . . . . 16 7.1 Resistance measurement by RCL multimeter, resistivity calcu-

lated afterwards by formula (1) . . . . 40 7.2 Resistance measurement by RCL Meter, resistivity calculated af-

terwards by formula (1) . . . . 40 7.3 Resistance measurement by digital multimeter, resistivity calcu-

lated afterwards by formula (1) . . . . 40 7.4 Resistance measurement by RCL multimeter in 37 ^◦ C, resistivity

calculated afterwards by formula (1) . . . . 41 7.5 Resistance measured by multimeter on the topside . . . . 41 7.6 Resistance measured by simple multimeter on the bottom side . 41 7.7 Bulk resistance measurement through the thickness of the sam-

ples using Galinstan droplet and Cu-tape . . . . 43

(6)

Abstract

A method of synthesizing silicone-based composite consisting of carbon black (CB) as a conductive filler in Polydimethylsiloxane (PDMS) was developed.

The aim was to find a cost effective and easier method to fabricate stretchable, epidermal and conductive electrodes in striving for inexpensive real-time health monitoring. In this work, instead of expensive additive materials for enhance- ment of PDMS conductivity, CB powder, at lower cost was used. To optimize the electrophysiological properties of the electrodes, limited amount of silver (Ag) and silver chloride (AgCl) particles were added. The electrical character- istic of the electrodes and their stretchability was studied.

Since fabrication and characterization did not require clean room enviroment,

the developed method was less costly and less time consuming. Samples were

made of six different filler concentrations in three sets, which in total were 18

samples, in order to obtain better statistics. Resistance of all samples was mea-

sured and resistivity values were calculated. Tensile test were performed on all

samples. The result showed that all samples had elongation of over 50 %, which

is feasible for stretchable, epidermal patches. Samples with filler concentration

of 10 wt% CB + 5 wt% Ag/AgCl and 10 wt% CB + 8 wt% Ag/AgCl showed

resistivity of Ωcm range. The electrodes were conductive, soft, stretchable and

biocompatible. They fulfill the requirements of epidermal patches for health

monitoring.

(7)

Chapter 1

Introduction

1.1 Stretchable Electronics

Flexible and stretchable electronics have applications in wearable and implantable electronic devices and have been a hot research subject for some time now.

Bendable, twistable and, most challenging, stretchable devices with their po- tential applications in medicine, energy and military can dramatically change the quality of human life.

Long term healthcare monitoring can become much more convenient because of their much lighter weight, softness and compliance. By using wearable sensors, for instance the patients’ physiological states can be monitored remotely and conveniently in real time.

Stretchable devices can be interfaced with non-planar, dynamic surfaces such as the human body, and remain reliable while subjected to strain.

Finding suitable material for different components of this kind of devices has been a crucial demand. These components need to be flexible, stretchable and soft, and depending on their application may need to have additional character- istics, such as being biocompatible and conductive, e.g., in case of implantable or epidermal electrodes [6], [7].

There are generally two main methods to fabricate stretchable electronics. One

method is based on using stretchable substances such as polymers that have

rubber-like behavior. The other method is to achieve a stretchable structure de-

spite, using non-stretchable materials, through a specific flexible design. As an

example, growing rigid semiconductor materials like silicon (Si) on prestretched

plane, then let it release. The result is, formation of buckling waves. For more

complex structures, origami (folding), and kirigami (cutting) techniques can be

used to make foldable, twistable or stretchable electronic devices Figure 1.1.

(8)

by using liquid alloys.

The advantage of using stretchable materials is that they can stand large elastic deformations due to their stretchability. Nevertheless, most stretchable mate- rials have high electrical resistance. Even though, rigid conductive or semicon- ductive material can be used in the second approach, more complicated designs are needed to achieve flexibility and/or stretchability in one or more dimensions [8].

Figure 1.1. Origami, a) 3-D sub-microscale pop-up silicon structure assembled by compressive buckling. b) Graphene origami by thermoresponsive self-folding.

c) Graphene kirigami by external force and actuation [1].

1.2 Epidermal Electrodes

Monitoring physiological processes is important for biomedical applications. By

utilizing electroencephalography (EEG), electrocardiography (ECG) and elec-

tromyography (EMG) to monitor brain, heart and skeletal muscles electrical ac-

tivities, more accurate diagnostics are possible. Traditionally, wired electrodes

have been used to collect and transfer data from the human body. Attaching

several electrodes to the body restricts normal body movements. To address

these issues, researchers came up with the wearable electronics that are in inti-

mate contact with skin and mimic its properties. Since then, a lot of effort has

been made to optimize these devices [9].

(9)

As the target for epidermal devices is to be attached to the skin for monitoring signals, some features are essential to be considered. Softness, stretchability and easy implementation of the epidermal electronics in wearable healthcare applications are those essential factors, which needs to be achieved for the near future. One of the commonly used materials in the approach to this technology is the soft, stretchable and biocompatible PDMS-based elastomer, that allows for making convenient patches with formation ability and letting the epidermal electronics be implemented in wearable applications [10].

1.2.1 Important Accomplishments and Concerns

1.2.1.1 Tattoo-like epidermal sensors

One example of wearable epidermal electronics is tattoo-like epidermal sensors Figure 1.2.

The graphene electronic tattoo (GET) can be attached on human skin directly, similar to temporary tattoos. The bare ones stay for a few hours but this time can be extended to a few days by using liquid bandage coverage. It can be peeled off from the skin by using adhesive tape. Their total thickness is 463 ± 30 nm, optical transparency of 85 % and stretchability of more than 40 %. These sensors have been used for ECG, EMG, EEG, skin temperature and hydration measurements. The main drawback is that some stages of the fabrication process need to be done mainly in a clean room such as graphene fabrication by atmospheric pressure chemical vapor deposition (APCVD) on a copper foil, spin-coating liquid polyimide on graphene and etching copper away.

Overall, this is an expensive and time consuming process [2].

Figure 1.2. Tattoo-like epidermal patch [2]

There are other materials like carbon fibers, metal films, silicon membrane and nanoparticle printable ink, which have been used in these kind of epidermal sensors, but here the ones made of graphene are briefly discussed.

1.2.1.2 PDMS-based Conductive Composites

(10)

conductivity of 10 ^–2 S.m ^–1 is achievable. The resistivity of this composite changes when changing the temperature. In range of 25 ^◦ C to 120 ^◦ C, the resistivity increases and reaches a peak, and after that it decreases. This char- acteristic, variable resistivity, can be useful for example in thermal sensors [11].

Other examples of conductive fillers that have been used to reinforce electri- cal conductivity in insulator polymers are carbon nanotubes and CB powder.

Carbon nanotubes have shown electrical conductivity higher than copper, which makes them a good candidate as conductive filler. On the other hand since their elongation to failure is in range of 20 % - 40 %, they are not reliable to sustain larger strain in tension, which can be a disadvantage for applications which de- mand higher elongation [12]. Carbon black PDMS (cPDMS) has shown lower percolation threshold in comparison with AgPDMS. This has been reported to be above 10 wt%. However, it has also shown the lower conductivity about 3 orders of magnitude less than AgPDMS according to Xize Niu et al [11].

1.3 Aim

The aim of this thesis was to fabricate and characterize soft, stretchable and

conductive electrodes from PDMS and CB composites. These electrodes are

targeted to epidermal patches, which eventually can be used in ECG wearable

sensors for healthcare monitoring. The important part of this project is to

achieve the goal by using low-cost, non-toxic materials and an easy method of

synthesis, which are important factors in comparison with other previous works

such as tattoo-like epidermal sensors. Unlike tattoo-like epidermal sensors, all

the process in this project was done in a chemistry lab, without any need of

being in a clean room. Moreover, using CB instead of other conductive fillers

such as silver or carbon nanotubes was more cost efficient.

(11)

Chapter 2

Theory

2.1 Principle of Epidermal Electronics

2.1.1 Relevant Skin Characteristics

Skin is the first barrier between the external environment and internal organs.

Wearable, non-invasive sensors should one way or another attached to the out-

ermost layer of skin, which is epidermis. Epidermis in turn consists of several

layers. Stratum corneum is its top layer with 10-100 μm thickness and is known

as skin barrier. This layer is dry (too little water) and oily therefore, electrically

resistive. The resistance of the stratum corneum is about 10 ⁵ Ω per cm ² and its

capacitance is 30 nFcm ^–2 . Moreover, epidermis is soft, stretchable, and covers

the underlying organs and consequently dampening the effects of mechanical

forces inside the body. All these reasons make it to be considered more of an

information barrier than the information source, when it comes to wearable

sensing [8]. It has been claimed that for stretchable electronics: “One of the

most attractive targets is for the human body, which endures strains of 30 % at

skins and over 100 % at joints” [13], [14].

(12)

Figure 2.1. Layers of skin [3]

2.1.2 Principle of Epidermal Electronic Sensor

Electrodes in wearable sensors function as a transducer to convert natural ionic flows in the body to measurable electrical signals. This can happen at electrode- electrolyte interface. Considering that there are no free ions in the electrodes and no free electrons in the electrolyte, chemical reactions need to happen to provide them at the interface. These reactions can be described as:

C−→C ⁺ + e ^– A ^– −→A+e ^–

where C and A are cations and anions, respectively, in the electrolyte. These reactions lead to presence of free electrons in the electrode, i.e., having an elec- tric current.

Body ionic flow at the interface disturbs the neutrality that result in the poten- tial build up. This potential difference is called half-cell potential, which can be different depending on what materials the electrodes are made of.

Biopotential electrodes can be polarizable or non-polarizable. Polarizable elec- trodes function as capacitors, which means no charge actually passes through the interface between the electrode and electrolyte. The current is because of displacement current, while in non-polarizable ones, current can pass across the electrode-electrolyte interface, which can be an advantage for these types of electrodes.

Generally, there are three different biopotential electrodes, dry, wet and non- contact electrodes [15].

In wet electrodes, gel is used as the electrolyte in between skin and electrodes.

The electrical impedance of the electrode-skin interface determines the quality

(13)

and efficiency of the recording. ‘Wet’ electrode contact is known as the best interface, which can be obtained using hydrogel or an electrically conductive adhesive on top of the electrodes [16].

In dry electrodes, no gel is used as contact medium between the electrode and the skin. These electrodes are more convenient, on the other hand, they do not have any electrolyte, which makes them more similar to polarized electrodes (capacitor).

Non-contact electrodes as their name suggests function as ”remote” sensors, that can detect the biopotential signals [16].

2.2 Percolation Threshold

Adding conductive fillers to polymer matrices to enhance their electrical or ther- mal properties is not something new. The principle is, adding enough amount of conductive fillers to create a pathway of conductive particles that can let the electric current pass through them. This will happen through two mechanisms, one is when the conductive particles are in contact with each other electrons can transfer through them, and another is when these particles are not con- nected but are close enough that quantum tunneling can happen. The lowest concentration of fillers at which the first continuous, conductive network forms is known as the percolation threshold. Below this amount, the inter-particular space is not small enough to let the electrons jump over and above this amount, loading the polymer matrix with higher filler concentration does not signifi- cantly improve the conductivity [4].

The percolation threshold is dependent on both fillers and polymers charac-

teristics and also their interactions with each other. It varies for different filler

materials, dimensions, shapes and segregations. For example, the aspect ratio of

the conductive particles is one important factor. The higher the aspect ratio, the

lower the percolation threshold [13]. Considering polymers’ molecular weight,

surface tension and crystallinity are determining factors. Polymers which have

higher molecular weight would have higher percolation threshold due to diffi-

culty in filler dispersion, while higher melt flow index and higher crystallinity

lead to lower percolation concentration. Higher surface tension, on the other

hand, increases percolation concentration which means decreasing conductiv-

ity. Furthermore, filler-filler interactions and filler-polymer interactions need

to be balanced, since if the first one is dominant then fillers will agglomerate

and percolation threshold will increase, and if filler-polymer interaction is dom-

inated then the inter-particular distances between fillers would be bigger and

consequently percolation concentration increases [17].

(14)

Chapter 3

Material

3.1 Elastomer

Since the electrodes in this project are supposed to be attached to the human skin, stretchability and softness were essential factors, which make them more comfortable.

PDMS fulfills all the requirements as both the substrate and elastic material for epidermal electrodes due to its chemical and physical properties such as bio- compatibility, flexibility, stretchability, non-toxicity and stability over a wide range of temperature from -50 ^◦ C to 200 ^◦ C. These properties of PDMS have made it popular in microelectromechanical system (MEMS) and in many appli- cations in industry and medical care [7].

There are a wide variety of silicone-based polymers in industry, but in this project, three of them were tested, which are listed in the table 3.1. The mix- ing ratio of elastomer parts is shown by (A:B), part A as base and part B as catalyst.

Table 3.1. Properties of elastomers used in this paper

Polymer Label Viscosity [mPa.s] Mixing Ratio-A:B Alongation at break [%] Cured form

Elastosil Vario15 Sigma-Aldrich 5000 10:01 900 Solid

Sylgard 184 Sigma-Aldrich 5500 10:01 140 Solid

Sylgard 527 Sigma-Aldrich 450 01:01 — Gel

Although PDMS is an insulator, by mixing conductive particles, conductivity

can be reached. In previous studies, researchers have investigated adding fillers

such as silver or carbon in different forms and shapes such as flakes, spheres,

tubes and fibers at various sizes from nano- to micrometer.

(15)

3.2 Fillers

3.2.1 Carbon Black

Carbon black was chosen as a filler, due to its cost effectiveness and good con- ductivity (ranging from 1 to 10 ⁴ S/m). Since lower cost production can mean higher accessibility of health care monitoring and therapeutic methods for ev- eryone, CB is a good candidate to be used in this application. CB appears as fine, black powder. It is amorphous in form of colloidal particles and its struc- ture resembles disordered graphite [18].

Carbon black primary particles are made of imperfect graphite layers. They are in the range of 10 to 500 nm with a surface area between 25 and 1500 m ² /g, which varies with their preparation method. Primary particles tend to fuse and form aggregates that are usually less than 1 μm. Usually 10 to 1000 of aggre- gates may join with each other and form clusters, which are called agglomerates [19].

Furnace black, thermal black, lamp black, channel black, and acetylene black are common CBs, with their various physical and chemical properties due to their different methods of production. Carbon black is 2-dimensional and its structure has been demonstrated in Figure3.1 [20].

Carbon black structure (high or low), refers to the number of primary parti- cles that have formed the aggregations and consequently affects the degree of dispersion and electrical conductivity [21], [4]. Apart from CB structure, other important factors are, the surface area and primary particle size. These factors influence the interaction between the particles and their medium.

For example CB structure affects the dispersion of particles because the more

they tend to agglomerate, the more difficult the dispersion. Therefore, CB

structure influences the electrical conductivity of the composite. As another

example, the ability of CB to absorb the UV radiation depends on its surface

area.

(16)

Figure 3.1. Carbon black primary, aggregates and agglomerations structure [4]

Polymer and CB interaction is influenced by the surface tension of the polymers and CB characteristics such as the ones mentioned above. To reinforced conduc- tivity to insulator polymer, it is needed to add sufficient amount of conductive filler to reach the percolation threshold. Also, even this amount depends on both polymer and CB properties and structures. According to Kausar 2018, the amount of CB that is needed to reach the percolation threshold varies be- tween 5 wt% and 20 wt% in the polymer matrix [13].

There are different kinds of carbon blacks with different grain shapes and sizes as mentioned before and in this work C65 was used. Carbon black grade numbers are based on surface area and structure, Figure 3.2.

(a) (b)

Figure 3.2. C65 TEM pictures (a) CB primary particles size 40 nm in aggre-

gates, (b) CB agglomerations [5].

(17)

3.2.2 Ag/AgCl

Ag/AgCl is a common biopotential electrode. Its advantages are:

1) Stability, when is in contact with biological fluids, which has high concentra- tion of Cl ^– .

Ag↔Ag ⁺ + e ^– Ag ⁺ + Cl ^– ↔AgCl↓

2) Non-polarizable, which means current passes freely through the interface of electrode and electrolyte. In addition non-polarizable electrodes have mini- mum motion artifacts.

3) Ag/AgCl electrodes have the lowest half-cell potential about 220 mV [16], [22], [23].

3.3 Dilutant

PDMS is a viscous material and mixing fillers with it is a challenge. Dilutant can be added to the mixture, to reduce the viscosity and make the mixing process easier. Three different sorts of dilutants were tested, ethanol, acetone and hexane. Some properties of these dilutants are shown in table 3.2.

Table 3.2. Dilutant properties

Dilutant Lable Chemical Formula Density

[ g/cm³]

Evapn.residue [%]

Boiling point [^◦C]

Acetonitrile Sigma-Aldrich CH 3 CN 0.786 <0.0005 81-82

Ethanol absolute VWR CHEMICAL C2H5OH 0.7895 78.3

Hexane Sigma-Aldrich CH3(CH2)4CH3 0.659 <0.0003 68.5-69.1

(18)

Chapter 4

Experimental Process

4.1 Method for Synthesis Composite

4.1.1 Mixing process:

First, it was started with sifting CB (C65, TIMCAL, aggregates of 200 nm) par- ticles with an ordinary (commercial) tea sift. Since CB particles were agglomer- ated this step was used to reduce the size of agglomerations and separating them.

Next step was weighing all the components one by one according to the desired weight ratio. Then, adding hexane as a dilutant to the fillers (Ag flakes<20μm;

Alfa Aesar by Thermofisher), AgCl (Acros Organic) and start mixing them on a magnet stirred for one hour. To slow down hexane evaporation, the container was sealed with paraffin film during the mixing process.

After an hour, the base part of Vario 15 (part A) was added. The mixing process was done by a hand-made electric vacuum mixer, made by Professor Hugo Nguyen at Uppsala University, Figure 4.1. This device will be referred as HD in the text. The HD device made it possible to mix and apply vacuum at the same time or separately. This stage also took about an hour to just mix the components without applying vacuum.

Next, Vario 15 catalyst (part B) was added. At this stage, all the compo- nents were mixed for 15 minutes, this time with applying vacuum .

The mixing process is shown in Figure 4.2.

(19)

Figure 4.1. Electric vacuum and mixer (HD device)

Figure 4.2. Mixing process

(20)

Afterwards, the mixture was poured in the handmade molds with rectangle shape that is shown in figure 4.3

Figure 4.3. The mold, which used while curing the samples

A spatula was used to make the surfaces of the poured material in the mold as smooth as possible, then they were kept under fume hood for 24 hours to be cured. After curing, removing samples from the molds and cut them with laser cutter (Ostling, AIO G+532 nm, 5W, and parameters set as power 55 %, shot frequency 1300, scan speed 200 and passes 220) in a desired size (Length: 50 mm, Width: 5 mm). Afterwards, resistance was measured for all samples and resistivity was calculated.

Three different devices were used to measure resistance, RCL meter, Key-Sight digital multimeter and simple digital multimeter, Figure 4.4. These devices were used either because of the different contacts that were compatible with them (like needle electrodes or flat clamps) or to have a more clear view over the samples’ resistance behaviors.

Samples with low filler concentrations showed a lot of fluctuations in resistance measurements, so the resistance was read after 5 seconds.

(a) (b) (c)

Figure 4.4. Three devices to measure resistance (a) RCL multimeter, (b)

Key-Sight multimeter, (c) Simple digital multimeter

(21)

4.2 Resistivity Calculation

Materials can be categorized according to their ability to resist or conduct elec- trical current. Conductors conduct electrical current fluently. Insulators do not conduct electricity at all. Semiconductors are conductive when receiving enough energy for example by heating them or when they are doped so that their prop- erties can be tuned. The resistivity (ρ) in Ohms.meter (Ωm) or Ohms.centimeter (Ωcm) was calculated by Formula (1):

ρ= R (A/L) (1)

where:

R is the electrical resistance in ohms(Ω).

L is the length of the piece of material in metres (m) or centimeter (cm).

A is the cross-sectional area in m ² or cm ² (width*thickness).

Resistivity is constant for each specific material and is a material property [18].

The conductivity (σ), which shows how good a material can transfer electric current will be calculated by Formula (2), and is in range of zero for insula- tors and to infinity for superconductors. The unit is 1/Ωm or 1/Ωcm, which is known as Siemens per metre or Siemens per centimetre, respectively.

σ=1/ρ (2)

(22)

Chapter 5

Results and Discussions

5.1 PDMS Choice

During this project, three different kinds of PDMS were tested, Sylgard 184, Sylgard 527, and Elastosil Vario 15. Among them the last one was chosen as the best candidate to fulfill the project purposes since the lower viscosity made the mixing and synthesis process much easier. This is because adding CB par- ticles to elastomers increases the mixture viscosity, which in turns makes the dispersion of the particles more difficult. In addition, its stretchability was much better in comparison with Sylgard 184.

The other advantage of Vario 15 for this work was that Vario 15 was easy to handle, in contrast to gels (the Sylgards), which were too sticky and could not be removed easily from the molds. Another disadvantage of gels in this project was that, since they were not sustaining their shapes, they were difficult to work with.

5.2 Dilutant Choice

Ethanol, hexane and acetone were used and tested. The difficulty of using ace- tone was that there were two different phases, liquid acetone and the composites of the materials with high viscosity. But the mixing process of fillers in PDMS became easier in comparison with using no dilutant.

Ethanol was difficult to remove after mixing stage because it did not evapo- rate even after using hot plate to heat the mixture. The composition never cured, when ethanol was used.

Hexane (Sigma-Aldrich) was chosen because it evaporates in room tempera-

ture and comparatively in a shorter time. It means after mixing stages and

when curing the samples, it would not be needed to go through any extra pro-

(23)

cess to get rid of the dilutant. All the hexane had evaporated after samples were cured.

5.3 Mixing Method Choice

Different methods of mixing were tested:

Hand mixing by using a spoon, magnetic stirring, ultrasonic bath, Dremel and handmade electric vacuum mixer (HD).

The first method tested in this paper was only mixing in beaker with a spoon by hand for long time more than 2 hours and without adding any dilutant. The result was not satisfactory, because there were a lot of CB clumps caused by agglomeration of carbon particles, and bubbles which appeared because of mix- ing composites by spoon. It was impossible to disperse the CB clumps and get rid of the bubbles just by mixing with this method even if before pouring the composites were vacuum pumped. Another problem was high viscosity because no dilutant was added. All these problems made it impossible to have smooth surface of samples by using spatula.

The other tested method was adding dilutant and then using magnet stirring and ultrasonic bath, but still there was problem with CB agglomerations and so not having a smooth surface or homogeneous composition.

Using two syringes connected to each other was another method for mixing.

Even if the dilutant was used but still mixing was not good and there were a lot of bubbles and CB agglomerations in the composites Figure 5.1.

(a)

(b)

(24)

The other trial was adding dilutant and using ultrasonic bath, magnet stir- ring and Dremel device, but still in compare with the last chosen methods this method was not good because still a lot of large CB agglomerations and a lot of bubbles, which could be seen even when the composite had been pouring in molds and after curing they could be seen as cracks.

In all of these methods, vacuum was applied directly after mixing and before pouring the composites in molds but still a lot of bubbles were in the samples.

After testing these different methods, the best results were shown when first, magnet stirring and then HD device was used. Otherwise, small air bubbles which were introduced to the mixture while mixing outside vacuum chamber were very difficult to remove by applying vacuum afterwards. The other ad- vantage of this method was that the less clumps and agglomerations could be seen.

5.4 Filler Concentration Choice

All the samples are made with the method described in section 4.1. Different weight percentages of CB powder were used and the maximum value that could be reached, was 13 wt%. Above this value, the samples cracked. The cracked samples were not usable for the application because it was not possible to even remove them from the mold as shown in the Figure 5.2.

Figure 5.2. Cracked Sample of 14 wt% carbon black still in their mold

As a result from CB PDMS (cPDMS), 13 wt% CB was the highest concentration

which could be reached. In this step, the goal was to keep the highest weight

percentage of CB while adding silver and silver chloride. The percentage ratio

amount of Ag/AgCl was 80 wt%/20 wt%.

(25)

The highest amount of CB and Ag/AgCl fillers in PDMS could reach 10 wt%

CB and 8 wt% Ag/AgCl.

Even though the amount of fillers reached 18 wt% with respect to the composite, over 50 % stretchability was achieved in all the samples, which was desirable stretchability for this application. This was only achievable if CB concentration does not exceed 10 wt%.

5.5 Tensile Test

All the samples’ stretchability have been tested using a ruler as shown in Fig- ure 5.3. Dog-bone shaped samples were cured in a mold Figure 5.3 (a), instead of being cut either by laser or hand from rectangular samples after being cured.

The reason was to prevent possible formation of micro-cracks. Samples were stretched slowly by hand. The length of the narrow part was measured both before and after stretching. Then by using Formula (3),

ε = Δl

l ₀ (3)

their stretchability has been calculated. For example, for the sample in the Figure 5.3 unstretched sample has an initial length of l ₀ = 3 cm and after applying tensile test the Δl = 6 cm so ε = 2, which means 200 % stretchibility.

All the samples have shown stretchability over 50 %. Samples with the lower

filler concentration reached over 100 % elongation. For epidermal electrodes,

any value above 50 % is sufficient.

(26)

(a)

(b)

(c)

Figure 5.3. Tensile test for sample 8 wt% CB + 5 wt% Ag/AgCl using ruler and stretching by hand. (a) Dog-bone mold, (b) Unstretched sample, (c) Stretched sample

5.6 Measurements and Implications

Three samples were made for each filler concentration. Resistance values were

measured and resistivity of each sample was calculated according to their di-

mensions. For each group of samples with the same filler concentration, mean

values of resistivity were calculated.

(27)

Figure 5.4. Average of measured resistivity of 3 series of samples

As it is shown in the Figure 5.4, taking the measurement error into account, increasing the filler concentration from 8 wt% CB to 10 wt% CB (columns 1 and 3, respectively) does not cause change of the composite resistivity, which is an indication that at this amount we have not reached the percolation thresh- old, since when a complete conductive pathway forms a drastic fall of resistivity value is expected.

Adding silver flakes to the composite in the case of 8 wt% CB + 5 wt% Ag/AgCl increased the resistivity, which means Ag/AgCl particles did not incorporate in forming a conductive pathway. Assuming that the critical concentration had not been achieved in the composite, dominant process for transferring the charge carriers is electron tunneling. One can assume that probably this process was interrupted by adding Ag/AgCl to the composite. One assumption can be that the filler-filler interaction between CB and Ag particles occurs in a way that increases the agglomeration of conductive particles. It may also be vice versa that this interaction increases the dispersion of CB particles too much. Since both of these two phenomena can increase the distance between the conductive particles, which decreases exponentially the probability of electron tunneling.

Another possibility is that AgCl particles, which are insulators, can come be-

tween CB and Ag particles and break the network between them causing the

resistivity to increase. On the other hand, when there are more Ag/AgCl par-

ticles available, as in the case of 10 wt% CB + 5 wt% or 8 wt% Ag/AgCl, and

since the ratio between Ag and AgCl is 80 % to 20 %, the excessive Ag particles

may compensate the defects induced by AgCl particles. Even though, dominant

species in forming the conductive network in the samples are CB particles due

(28)

and facilitate the electron transfer[24].

To study the resistivity change in electrodes when they are mounted on the skin and therefore exposed to the temperature (in ^◦ C) higher than room temperature typically about (37 ^◦ C), hot plate was used to reach the wanted temperature in electrodes and to see how the samples’ behavior changes. Samples resistance was measured once in room temperature and another time while they were on top of the hot plate for 5 minutes. The temperature of hot plate was measured by infrared camera (FLIR) and digital thermometer (center 309 DATA LOG- GER). The results are shown in Figure 5.5.

Figure 5.5. Temperature dependency of resistivity

It can be observed that resistivity increased in the samples when temperature increased.

To find out whether the electrical resistivity of the samples has a directional components and a gradient along the thickness of the samples, sheet resistance on the top side and bottom side of the samples was measured by simple digital multimeter.

The comparison between measurement results obtained from each side for 3

batches is shown in the Figure 5.6.

(29)

(a)

(b)

(c)

(30)

The amount of fillers influences the viscosity, namely, higher amount of fillers makes the PDMS more viscous. CB dispersion morphology differs correspond to viscosity of the composite during the curing time. When viscosity is low the filler particles sink easily and it will affect electrical conductivity. It is seen in the Figure 5.6 for lower filler concentrations like 8 wt% CB and 8 wt% CB +5 wt% Ag/AgCl the sheet resistance is higher on the top side than that on the bottom side. This is not true for the samples with the higher amount of fillers such as 10 wt% CB + 3/5/8 wt% Ag/AgCl since they showed the same values of sheet resistance . On the other hand, if it was a possibility to decrease the curing time by curing the composite in oven and not in room temperature, so that the conductive particles might not have enough time to sink and the dif- ference between resistivity of bottom and top side might have been less.

In other words, the electrical resistivity of the composites with CB filler con- centration above 10 wt% seems less affected by sedimetation of filler particles and as a result there was smaller difference between measured resistance values on their top and bottom side.

It was observed that the amount of CB fillers in comparison to Ag fillers, in- fluences more the composite viscosity. This is due to lower density of CB than that of Ag, which means the volume-% of CB is much larger than that of silver for the same mass [25]. Since ideally this material is going to be used on the skin as the electrodes, it is important to know how much resistance will oppose the current passing through the bulk thickness. Moreover, cPDMS is soft and stretchable so using clamps that were available means compressing the material, which theoretically can affect the conductivity.

To study this effect, the set up was used, which is shown in Figure 5.7 for mea-

suring the resistance through the thickness of the samples. This time, instead

of clamps, needle electrodes were used. The sample was sandwiched between a

layer of Copper (Cu) tape and a droplet of Galinstan. Cu-tape was fixed on a

glass substrate, while the cPDMS sample was placed on top of it in a way that

part of Cu-tape was not covered by the sample. This part was then used for

probing by one of the electrodes. Then, a droplet of Galinstan was placed on top

of the sample, where the second electrode later was placed to measure the re-

sistivity through the thickness of the sample. The result is shown in Figure 5.8.

(31)

(a)

(b)

Figure 5.7. Bulk resistance measurement with Galinstan droplet and Cu-

tape, (a) Schematic view of measurement setup, (b) Resistance measurement

by simple digital multimeter

(32)

Figure 5.8. Average of measured resistivity of batches of 3 samples

(33)

Comparing average resistivity results in Figure 5.4 and Figure 5.8, shows a dras- tic difference in 3 orders of magnitude. The resistivity values are in range of Ωcm in the first graph and kΩcm in the second one.

It should be noted that resistivity values demonstrated in Figure 5.4, were cal- culated based on the lengthwise resistance measurements, while the values in Figure 5.8, were measured across the thickness of the samples. This becomes important because the path that electric current needs to pass is different in these two cases.

The way that conductive fillers sit beside each other in two different axis influ- ences the measured resistance and consequently resistivity values.

As it was mentioned before, because of the particles’ sedimentation, fillers are not homogeneously dispersed in the samples. This inhomogeneity was much less pronounced for the samples with the higher filler concentrations due to the less viscosity that composites had before curing.

In higher concentrations of fillers, sedimentation of particles decreases, at the same time there are more conductive particles in the whole sample. When mea- sured lengthwise the current passes through the ”layer”, which has the higher concentration of conductive particles. Whereas, when measured across the thick- ness, the current should pass through the whole sample from the top side with fewer conductive particles to the bottom, which has more conductive particles.

That can be one possible reason that resistance through the thickness were in- creased.

The sedimentation and, as a result, the inhomogeneity in samples makes this huge difference between the measured resistance values along horizontal and vertical axis.

Figure 5.9 shows some overlaps concerning the resistivity results of different filler concentrations. This suggests the possibility, that the results cannot be considered different and the small difference in the obtained values was actually because of measurement errors.

To study this possibility, a statistical method was used, which is known as

”student’s T-test”.The overlapped intervals are shown in Figure 5.9. Black dots show the actual obtained values and black holes show the standard deviation.

As an example, if we compare the result intervals from 8 wt% CB with the

ones belong to 10 wt% CB and 10 wt% CB + 3 wt% Ag/AgCl, an obvious

overlap is visible, which means these values can be considered the same and we

cannot see clearly the difference between them. However, 8 wt% CB + 5 wt%

(34)

Figure 5.9. Students’ T-test (statistical test) to check the range of possible

results, the black dots are resistivity experimental values for different samples

with the same compositions, black circles are mean values and errorbars are

showing the confidence intervals that is the range, where the results are probable

to occur as high as 95%.

(35)

Chapter 6

Conclusion

A method of synthesizing CB and Ag/AgCl PDMS has been developed for use as stretchable electordes or electrical conductor.

Concentration of CB and Ag/AgCl in PDMS was varied to study and optimise the electrical conductivity in PDMS. Carbon black was used as a conductor and Ag/AgCl was used for converting ionized current from body to electric current.

The highest single filler loading concentration of CB in PDMS achieved was 13 wt %. The highest concentration of Ag/AgCl, that was added to 10 wt% CB was 8 wt%.

Fabrication process was done in ordinary chemistry lab (avoiding high-cost of and time consuming in clean room environment).

After making sheets of the material, samples were cut by using laser cuter and their resistances were measured by using three different devices to establish a clear image of the samples’ electrical behavior. Then, their resistivity was calculated.

All samples showed stretchability over 50 % in tensile tests, which fulfils the requirement for using them as epidermal patches.

Resistivity of all the samples were investigated at room temperature and at 37 ^◦ C to predict the possible changes in the case they are attached on the skin.

It showed that resistivity of all samples increased as the temperature increased, with small differences for samples with higher filler concentration.

Measurements performed on samples with by using Galinstan droplet and Cu-

tape showed a dramatic increase in resistivity compared with the ones with

clamps, except for 10 wt% CB +5 wt% Ag/AgCl and 10 wt% CB + 8 wt%

(36)

Chapter 7

Outlook

Filler dispersion plays a very important role in determining the characteristics of the composite. Any improvement in mixing process can effect the filler dis- persion and the final result. Testing other methods of mixing such as using miller [11],melt blending or in-situ polymerization [4] is highly recommended.

As another suggestion adding surfactant to improve the affinity of the filler par- ticles to the elastomer matrix may ease the mixing process [14].

Another recommendation is, to cure the samples with faster methods for ex- ample curing them in the oven can be helpful. The faster the samples cure, the shorter time the fillers have to settle, which means more homogeneity in the matrix.

When samples are ready, it can be beneficial to use TEM or SEM to deter- mine the distribution of fillers in the bulk and surface, which can help clarifying electrical and mechanical behavior of the composite.

Determining CB structure (high or low), surface area and aggregation density because these factors can determine filler/filler and filler/polymer matrix inter- actions.

In general the fabrication process needs to be done more meticulously. For instance, it is needed to keep the fraction of dilutant volume percent to filler volume percent constant in order to keep the composition viscosity the same for all the samples. The viscosity is an important case because it can affect the rate of fillers sedimentation.

The samples thickness is also affected by the dilutant amount.

Another point to be considered during mixing process is temperature. Tem-

perature affects the rate of dilutant evaporation so it is important to have the

same temperature for all different compositions while mixing and curing them.

(37)

Investigating about interaction between CB and Ag particles in this case, which

there are more than one filler can be helpful to know the electrical behaviour

av samples better.

(38)

References

[1] Xin Ning, Xueju Wang, Yi Zhang, Xinge Yu, Dongwhi Choi, Ning Zheng, Dong Sung Kim, Yonggang Huang, Yihui Zhang, and John A Rogers. As- sembly of advanced materials into 3d functional structures by methods inspired by origami and kirigami: a review. Advanced Materials Interfaces, 5(13):1800284, 2018.

[2] Shideh Kabiri Ameri, Rebecca Ho, Hongwoo Jang, Li Tao, Youhua Wang, Liu Wang, David M Schnyer, Deji Akinwande, and Nanshu Lu. Graphene electronic tattoo sensors. ACS nano, 11(8):7634–7641, 2017.

[3] Melanocyte and melanin. layers of the human skin. URL https://www.shutterstock.com/image-illustration/

melanocyte-melanin-layers-human-skin-177142175?src=v_

7jcAa8EOlB2zMWQuRJAA-1-14.

[4] Wilailak Chanklin. Electrical properties study of carbon fillers in polymer nanocomposites. PhD thesis, University of New Brunswick, 2016.

[5] Peter Flanagan. personal communication, September 2019.

[6] Hao Wu, YongAn Huang, Feng Xu, Yongqing Duan, and Zhouping Yin.

Energy harvesters for wearable and stretchable electronics: from flexibility to stretchability. Advanced materials, 28(45):9881–9919, 2016.

[7] Hailin Cong and Tingrui Pan. Photopatternable conductive pdms materials for microfabrication. Advanced Functional Materials, 18(13):1912–1921, 2008.

[8] J. Heikenfeld, A. Jajack, J. Rogers, P. Gutruf, L. Tian, T. Pan, R. Li, M. Khine, J. Kim, J. Wang, and et al. Wearable sensors: modalities, challenges, and prospects, Nov 2017. URL https://pubs.rsc.org/en/

content/articlehtml/2018/lc/c7lc00914c.

[9] Milad Behbahaninia, Nikolay L Martirosyan, Joseph Georges, Joshua A Udovich, M Yashar S Kalani, Burt G Feuerstein, Peter Nakaji, Robert F Spetzler, and Mark C Preul. Intraoperative fluorescent imaging of intracra- nial tumors: A review. Clinical neurology and neurosurgery, 115(5):517–

528, 2013.

(39)

[10] Seung Hee Jeong, Shuo Zhang, Klas Hjort, J¨ ons Hilborn, and Zhigang Wu.

Pdms-based elastomer tuned soft, stretchable, and sticky for epidermal electronics. Advanced Materials, 28(28):5830–5836, 2016.

[11] XZ Niu, SL Peng, LY Liu, WJ Wen, and Ping Sheng. Characterizing and patterning of pdms-based conducting composites. Advanced Materials, 19 (18):2682–2686, 2007.

[12] GE Zaikov, GV Kozlov, Z Kh Aphashagova, and AK Mikitaev. Plastic- ity structural analysis of polymer nanocomposites filled with carbon nan- otubes. Journal of Applied Polymer Science, 111(5):2621–2624, 2009.

[13] Ayesha Kausar. Contemporary applications of carbon black-filled polymer composites: an overview of essential aspects. Journal of Plastic Film &

Sheeting, 34(3):256–299, 2018.

[14] Naoji Matsuhisa, Martin Kaltenbrunner, Tomoyuki Yokota, Hiroaki Jinno, Kazunori Kuribara, Tsuyoshi Sekitani, and Takao Someya. Printable elas- tic conductors with a high conductivity for electronic textile applications.

Nature communications, 6:7461, 2015.

[15] Refet Fırat Yazıcıo˘ glu, Chris Van Hoof, and Robert Puers. Introduction to biopotential acquisition. In Biopotential Readout Circuits for Portable Acquisition Systems, pages 5–19. Springer, 2009.

[16] Stephen Lee and John Kruse. Biopotential electrode sensors in ecg/eeg/emg systems. Analog Devices, 200:1–2, 2008.

[17] M Norkhairunnisa, A Azizan, M Mariatti, H Ismail, and LC Sim. Thermal stability and electrical behavior of polydimethylsiloxane nanocomposites with carbon nanotubes and carbon black fillers. Journal of Composite Materials, 46(8):903–910, 2012.

[18] Ayesha Kausar. Contemporary applications of carbon black-filled polymer composites: An overview of essential aspects - ayesha kausar, 2018. URL https://journals.sagepub.com/doi/10.1177/8756087917725773.

[19] Ann Y Watson and Peter A Valberg. Carbon black and soot: two different substances. AIHAJ-American Industrial Hygiene Association, 62(2):218–

228, 2001.

[20] R Sanjin´ es, Manuel D Abad, Cr Vˆ aju, R Smajda, M Mioni´ c, and A Ma- grez. Electrical properties and applications of carbon based nanocomposite materials: An overview. Surface and coatings technology, 206(4):727–733, 2011.

[21] Omar A Al-Hartomy, Falleh Al-Solamy, Ahmed Al-Ghamdi, Nikolay

Dishovsky, Milcho Ivanov, Mihail Mihaylov, and Farid El-Tantawy. Influ-

ence of carbon black structure and specific surface area on the mechanical

(40)

[22] Libretexts. Reference electrodes, Jun 2019. URL https://chem.

libretexts.org/Bookshelves/Analytical_Chemistry/Supplemental_

Modules_(Analytical_Chemistry)/Analytical_Sciences_Digital_

Library/JASDL/Courseware/Analytical_Electrochemistry:

_Potentiometry/03_Potentiometric_Theory/04_Reference_

Electrodes.

[23] John G Webster. Reducing motion artifacts and interference in biopotential recording. IEEE Transactions on Biomedical Engineering, (12):823–826, 1984.

[24] JL Acosta, MC Ojeda, and C Del R´ıo. Effect of conducting fillers on the microstructure and electrical conductivity of thermoplastic polymer composites. Polymer Bulletin, 57(2):199–206, 2006.

[25] S-P Rwei, F-H Ku, and K-C Cheng. Dispersion of carbon black in a con-

tinuous phase: Electrical, rheological, and morphological studies. Colloid

and Polymer Science, 280(12):1110–1115, 2002.

(41)

Appendix

Table 7.1. Resistance measurement by RCL multimeter, resistivity calculated afterwards by formula (1)

Resistivity(Ωcm) CB+Ag/AgCl

w% 8% 8%+5% 10% 10%+3% 10%+5% 10%+8%

Serie 1 Serie 2 Serie 3

7.2 6.6 8.7

12 11.6 10.2

8.8 10.8

4.3 5.9 6.4 7.8

4.8 5.7 3.6

5.2 4.8 6.3 RCL multimeter device, hand-cut samples(L:50mm, W=5mm)

Table 7.2. Resistance measurement by RCL Meter, resistivity calculated af- terwards by formula (1)

Resistivity(Ωcm) CB+Ag/AgCl

wt% 8% 8%+5% 10% 10%+3% 10%+5% 10%+8%

Serie 1 Serie 2 Serie 3

7.8 6.8 6.8

13.5 13 13.1

7.4 9.3 6.5

7.3 8.2 8.9

4.6 4.3 3.5

3.4 4.4 4.2 RCL multimeter device, Laser-cut samples (L:50mm, W=5mm)

Table 7.3. Resistance measurement by digital multimeter, resistivity calcu- lated afterwards by formula (1)

Resistivity(Ωcm) CB+Ag/AgCl

wt% 8% 8%+5% 10% 10%+3% 10%+5% 10%+8%

Serie 1 Serie 2 Serie 3

9.1 6.8 9.0

13.5 14.5 14.4

8.0 9.0 7.7

8.2 9.8 9.7

4.7 4.3 4.1

3.5

5.1

4.6 Key-sight digital multimeter, Laser-cut samples(L:50mm, W=5mm)

(42)

Table 7.4. Resistance measurement by RCL multimeter in 37 ^◦ C, resistivity calculated afterwards by formula (1)

Resistivity(Ωcm) CB+Ag/AgCl

wt% 8% 8%+5% 10% 10%+3% 10%+5% 10%+8%

Serie 1 Serie 2 Serie 3

9.6 9.7 7.4

17.1 15.3 15.3

8.6 10.0

7.1 8.7 9.8 10.1

5.1 4.8 3.8

3.8 4.7 4.6 RCL multimeter device, Laser-cut samples (L:50mm, W=5mm, T=37

^◦

C)

Table 7.5. Resistance measured by multimeter on the topside Top side sheet resistance(kΩ)

CB+Ag/AgCl

w% 8% 8%+5% 10% 10%+3% 10%+5% 10%+8%

Serie 1 Serie 2 Serie 3

3.4 1.6 2.5

6.8 4.6 4.8

2.3 2.7 1.6

2.7 3.0 3.0

1.8 1.4 1.6

1.6 1.8 1.6 Simple multimeter, Laser-cut samples(L:50mm, W=5mm)

Table 7.6. Resistance measured by simple multimeter on the bottom side Bottom side sheet resistance(kΩ)

CB+Ag/AgCl

w% 8% 8%+5% 10% 10%+3% 10%+5% 10%+8%

Serie 1 Serie 2 Serie 3

2.8 2.3 2.3

4.6 2.7 3.9

3.4 3.0 1.5

2.7 3.2 3.2

1.8 1.7 1.6

1.6

1.8

1.6 Simple multimeter, Laser-cut samples (L:50mm, W=5mm)

(43)

Figure 7.1. Reisistivity vs filler concentration,RCL multimeter, hand-cut sam- ples

Figure 7.2. Reisistivity vs filler concentration,Simple digital multimeter,

Laser-cut samples

(44)

Figure 7.3. Reisistivity vs filler concentration,RCL multimeter and hotplate, Laser-cut samples

Table 7.7. Bulk resistance measurement through the thickness of the samples using Galinstan droplet and Cu-tape

Resistivity (kΩcm) CB +Ag/AgCl

wt% 8% 8%+5% 10% 10%+3% 10%+5% 10%+8%

Serie 1 Thickness(cm)

3.7 0.06

4.9 0.04

9.6 0.05

4.5 0.04

6.7 0.04

0.84 0.04 Serie 2

Thickness(cm) 7.4 0.04

3.3 0.07

11.6 0.04

5.6 0.04

0.3 0.03

0.01 0.05 Serie 3

Thickness(cm) 2.6 0.04

2.6 0.05

4.2 0.05

3.3 0.03

0.01 0.03

0.1

0.04 Simple digital multimeter, Hand cut samples, Galinstan droplets, Cu-tape

Silicone-based Carbon Black Composite for Epidermal Electrodes