3
Experimental Setup

3.1 Introduction

M olecular communication, as discussed in the previous chapters, is an emerging field, and compared to traditional communication methods (i.e., EM ) is still in its infancy regarding the understanding of its physical properties. To analyse this novel communication paradigm experimentally, a test-bed must be established. Unlike EM or AC , molecular communication does not rely on the propagation of waves. Instead, it relies on the momentum of travelling particles. Therefore, to realise the communication, a receiver that can distinguish chemicals must be implemented. This can be accomplished in two ways.

The first method is to use sensors that physically interact with the particles and produce current from the interaction (i.e., MOSFET s). Consider, as an example, tin-oxide ( $S N O_{2}$ ) sensors. The gates of these MOSFET s are covered with $S N O_{2}$ and by interacting with the $O_{2}$ in the air it produces a current, depending on the concentration value of the introduced chemical species. However, this can cause a bottleneck in the communication where the interaction of the chemical and the sensor can decrease the sensitivity of the future transmissions and make it unsuitable for long transmissions. Due to the mechanics of $S N O_{2}$ (redox) the sensor can reach saturation point where the sensor lacks $O^{-}$ ions to interact with the signal chemical in the transmission medium [330].

The second approach, is the utilisation of mass spectrometers (MS). A MS is an instrument capable of analysing and distinguishing charged ions based on their motion in an applied electric and/or magnetic field. The analyser of the MS allows the detection of ions with a particular mass-to-charge (m/z) ratio [251], [331], making it a useful tool for use in molecular communication, as described in Section 2.10.2.2 .

In this chapter, different parts of the experimental setup are described in detail. Section 3.2 gives an overview of the experimental test bed. Section 3.2.1 discusses the operation of the transmitter. In Section 3.2.2 the chemical species used throughout the thesis are given. Section 3.2.3 is dedicated to the receiver, where the operating principle of the QMA is given in detail.

Figure 3.1: The diagram of the experimental setup: (1) (

N_{2}

) gas is used as the carrier flow (

Q

) and is transferred into the MFC that control both the carrier flow (blue line) (

Q

) and the signal flow (yellow line) (

q

), (2) Modulation information is generated using computer software, (3) generated modulation is transmitted into an automation platform where it sends the modulation to the MFC’s to create pulses, (4) MFC for the carrier flow, (5) MFC for the signal flow, (6) Evaporation Chamber (EC) where the signal chemical is injected, (7) Mixing chamber where the signal chemicals arrive and initiate the transmission from the transmitter to the receiver (8) Propagation medium (9) The inlet of the mass spectrometer, (10) Electronics control unit (ECU) which controls the mass analyser, (11) Pressure gauge, (12) Controller and the regulator cables for the mass spectrometer, (13) Data acquisition and analysis.

3.2 Experimental Setup

The generation and transmission of chemicals based on a message was made using an in-house-built odour generator, and the detection of the chemical was made using a mass spectrometer (MS), namely a quadrupole mass analyser (QMA). The diagram for the experimental setup used in the study can be seen in Figure 3.1 . In the following section, each of the major parts of the experimental study is discussed in detail.

3.2.1 Transmitter

A gas dispenser that is controlled by mass flow controllers (MFCs) releases volatile organic compounds (VOCs) (placed into an evaporation chamber) called the signal flow, whereas the flow passing through the mixing chamber is called the carrier flow ( $N_{2}$ ) [25], [332]. VOC s are chemicals that can transmit from the liquid to the gas phase at ordinary room temperature and pressure. This property is due to the low boiling point of the chemical, which forces a large number of molecules to evaporate from its liquid or sublimate from its solid phase and mix-in with the surrounding air, known as volatility. Most odours are made up of VOC s.

The MFC s are connected directly into the $N_{2}$ gas cylinder and controlled directly by a computer and an automation platform. The automation platform sends digital commands to the MFC and the internal valve inside can be controlled to create gas pulses. In this setup, there are two flows present: one is responsible for transporting the samples from the evaporation chamber into the mixing chamber (signal flow $q$ ); the second one is responsible for carrying the signal chemicals from the mixing chamber to the transmission medium (carrier flow $Q$ ). A diagram of the odour gas generator can be seen in Figure 3.2 , and the evaporation chamber can be seen in Figure 3.3 . When the gas leaves the mixing chamber and transmits through the medium it is defined as bulk flow ( $B_{F}$ ) [25], [332]:

B_{F} = Q + q .

(3.1)

Figure 3.2: The working diagram of the odour generator ( OG ). [ 332 ] (1) Introduction of the carrier gas (Q) into the mixing chamber. (2) Mixing chamber where the evaporated chemicals from the chamber and the carrier gas are mixed. (3) Evaporation chamber (Figure 3.3 ). (4) Transmitted chemicals that are released from the chamber. (5) A modulation sequence that is used to create gas pulses.

Figure 3.3: Diagram of the evaporation chamber ( EC ): (1) inlet of the

N_{2}

gas into the evaporation chamber. (2) Inlet where the sample is introduced. (3) Thermo-resistant septum that allows the multiple introduction of a sample via a micro syringe. (4) An absorptive material that holds the liquid sample analyte. (5)

N_{2}

from the inlet carries the evaporated chemicals from the chamber. (6) The cumulated gas is transferred into the mixing chamber via a 6.35 mm inch Teflon tube.

3.2.2 Chemicals

In this study four types of chemicals were used. A zero-grade $N_{2}$ (% 99.998 purity with 1 bar pressure) was chosen for the carrier gas ( $Q$ ) that carries both the signal chemical from the transmitter to the receiver and transports the signal chemical from the evaporation chamber to the transmitter. The properties of the signal gas ( $q$ ) used in this study are given in detail in Table 3.1 . The mass spectrum of the chemicals used throughout the thesis can be seen in Figure 3.4 .

Table 3.1: Chemicals used throughout the thesis.

Parameters	Acetone	Methanol	Cyclopentane	n-Hexane
Molecular Formula	$(C H_{3}) C O C H_{3}$	$(C H_{3}) O H$	$C_{5} H_{10}$	$C_{6} H_{14}$
Molecular Weight	58.08 g/mol	32.04 g/mol	70.135 g/mol	86.18 g/mol
CAS Number	67-64-1	67-56-1	287-92-3	110-54-3
Vapour Density ¹	1.56	1.11	2.4	2.97
Vapor Pressure ²	184 mmHg	97.68 mmHg	317.8 mmHg	153 mmHg
Boiling Point	56 ° C	64.7 ° C	49.2 ° C	68.5 ° C
Melting Point	-17.2 ° C	-98 ° C	-94 ° C	-96 ° C
Diffusivity in air	0.124 $c m^{2} ∕ s$	0.15 $c m^{2} ∕ s$	0.0919 $c m^{2} ∕ s$	0.2 $c m^{2} ∕ s$
Chapters used in	5, 6 and 7	5, 6 and 7	7	7

¹ Relative to the density of the air (air = 1) ² Normal Temperature and Pressure

Figure 3.4: Mass spectra of the signal chemicals used in the study. The mass spectra of the chemicals are obtained via the use of electron ionisation ( EI ) [NIST Chemisty Thesis (https://webbook.nist.gov/chemistry)].

3.2.3 Receiver

In the studies following the chapter, to detect the chemicals that are released by the odour generator (OG), a portable membrane-inlet mass spectrometer (MIMS), from Q Technologies Ltd., was used as the primary sensor. MIMS is a method of introducing analytes into the MS ’s vacuum chamber via a semi-permeable membrane [333], [334]. The membrane is usually a thin, gas-permeable, hydrophobic material such as polydimethylsiloxane (PDMS), which is also used in this study. Samples can be almost any fluid, including water, air or sometimes even solvents. The great advantage of the method of sample introduction is its simplicity. MIMS can be used to measure a variety of analytes in real-time, with little or no sample preparation. MIMS is most useful for the measurement of small, non-polar molecules (< 200 Da), since molecules of this type have a greater affinity for the membrane material than the sample.

A MIMS consists of three primary parts: a sampling probe that lets the gas sample pass the membrane for the MS to analyse, a triple filter quadrupole mass spectrometer, which in turn consists of an electron ionisation (EI) source, QMA and a detector, and finally a vacuum system. The inlet of the system, consists of a fine non-sterile flat PDMS membrane [25], [335]. In the following subsections 3.2.3.1 and 3.2.3.2 major parts of the MIMS are described in detail.

3.2.3.1 Membrane Inlet

The initial analysis of the chemical species begins with the interaction of the signal chemicals with the membrane present in the inlet of the QMA. This process, called pervaporation , involves three distinct actions [336].

1.: Absorption of sample molecules into the membrane (absorption),
2.: Diffusion of sample molecules through the membrane (diffusion),
3.: Evaporation of the molecules from membrane surface to the vacuum (desorption).

The pervaporation process can be described using Fick’s two laws of diffusion. The $1^{s t}$ law describes the relation between the diffusive flux ( $J (x, t)$ ) and the concentration of the molecules in the given environment. The diagram of the pervaporation phenomena can be seen in Figure 3.5 .:

J (x, t) = - D \frac{\partial c (x, t)}{\partial x},

(3.2)

and the concentration gradient can be explained by Fick’s $2^{n d}$ law :

\frac{\partial c (x, t)}{\partial t} = D \frac{\partial^{2} c (x, t)}{\partial x^{2}} .

(3.3)

Figure 3.5: Diagram that illustrates the process of pervaporation.

The applications and the properties of the MIMS are described in the literature [335], [337], [338], [339], [340], [341]. Due to the presence of the membrane, each chemical can interact with the membrane differently, and this in turn can cause different absorption rates [25]. Additional details of the membrane and its applications can be seen in [335].

Figure 3.6: A diagram of the inner workings of the QMA. The sample is introduced into the system via a conduit, in which case is the membrane inlet and the sample is then bombarded with electrons in the EI to ionise. The ionised particles are then transferred into the quadrupole area via the focus lens which is fed with a DC voltages. The ions then travel the duration of the path inside the quasi-static electric field generated via the quadrupole. Only the ions with the correct mass-to-charge ratio have stable trajectories within the field generated via RF/DC applied to the electrodes. These ions are safely transported from the initial introduction to the electron multiplier. Ions with unstable trajectories, however, collide with the rods, or the walls of the chamber and are neutralised. Each ion passing through the mass analyser has a stability profile which is related to the DC and RF voltages applied to the rods. By adjusting these voltage values only specific ions can be monitored. This process, which is used in this study, is single ion monitoring (SIM).

3.2.3.2 Quadrupole Mass Analyser

Developed at the beginning of the 1950s by Wolfgang Paul and Steinwegen from the University of Bonn, the QMA relies on the manipulation of ion trajectories by controlling the RF voltage applied to the quadrupole rods [342]. Since then, QMA s have become an important instrument in analysing samples. The principle theory of QMAs is given in the following paragraph and a functional diagram of the workings of a QMA can be seen in Figure 3.6 .

Ions travelling along the $z$ -axis (axial axis) are under the influence of an electric field generated by four parallel metal rods with their centre at the origin. The potential lines and the vectorial lines of the quadrupole field can be seen in Figure 3.7 .

Figure 3.7: (A) Equipotential lines generated by the quadrupoles (B) Electric vector field generated by the quadrupoles.

The potentials applied to the rods are given below:

\begin{align} + Φ_{0} = + (U - V \cos ω t), & (3.4a) \\ - Φ_{0} = - (U - V \cos ω t), & (3.4b) \end{align}

where $Φ_{0}$ defines the potential applied to the rods ( $V$ ), $ω$ is the angular frequency ( $r a d ∕ s$ ), $U$ is the direct potential ( $V$ ) and $V$ is the “ zero-to-peak” amplitude of the RF voltage ( $V$ ).

The accelerated ions enter the space between the quadruples, which align along the $z$ -axis. The force acting on the accelerating ions is caused by the quadrupole electric fields. The possible trajectories of the ion inside a quadrupole field can be seen in Figure 3.8 .

\begin{align} F_{x} & = m \frac{d^{2} x}{d t^{2}} = - z e \frac{\partial Φ}{\partial x} = - 2 z e x \frac{(U - V \cos ω t)}{r_{0}^{2}}, & (3.5a) \\ F_{y} & = m \frac{d^{2} y}{d t^{2}} = - z e \frac{\partial Φ}{\partial y} = + 2 z e y \frac{(U - V \cos ω t)}{r_{0}^{2}}, & (3.5b) \end{align}

where $m$ is the mass of an ion and $e$ is the electric charge of a single electron ( $C o u l o m b s$ ). The electric potential ( $Φ$ ) can be represented as a function of ( $Φ_{0}$ ):

Φ_{(x, y)} = (x^{2} - y^{2}) \frac{Φ_{0}}{r_{0}^{2}},

(3.6)

where $2 r_{0}$ is the closest distance between the four electrodes. By rearranging the equations given in Eq. ( 3.5 ) , the movement equations (aka., Paul equation) is obtained:

\begin{align} \frac{d^{2} x}{d t^{2}} + \frac{2 z e}{m r_{0}^{2}} (U - V \cos ω t) x = 0, & (3.7a) \\ \frac{d^{2} y}{d t^{2}} - \frac{2 z e}{m r_{0}^{2}} (U - V \cos ω t) y = 0 . & (3.7b) \end{align}

Figure 3.8: Examples of stable and unstable ion trajectories in a quadrupole.

The equations above describe the movement of an ion in a 3D environment within the influence of a quadrupole electric field. The ion’s trajectory will stay stable if the $x$ and $y$ values never reach $r_{0}$ . If the values exceed $r_{0}$ the ion will collide with the quadrupole and never arrive at the detector. To solve the equations with respect to $x$ and $y$ the Eq. ( 3.7 ) needs to be integrated. To solve these sets of equations, the following conversions are used:

ξ = \frac{ω t}{2}, ξ^{2} = \frac{ω^{2} t^{2}}{4} .

(3.8)

To simplify the equations given in ( 3.7 ), an umbrella term of $Γ_{m}$ is given with the following identity with the redefined equation:

\begin{align} Γ_{u} & = x = - y, & (3.9a) \\ \frac{d^{2} Γ_{u}}{d t^{2}} + \frac{2 z e}{m r_{0}^{2}} & (U - V \cos ω t) Γ_{u} = 0 . & (3.9b) \end{align}

Including the transformation given in Eq. ( 3.8 ) to Eq. ( 3.9 ) yields the following expression:

\frac{ω^{2}}{4} \frac{d^{2} Γ_{u}}{d ξ^{2}} + \frac{2 z e}{m r_{0}^{2}} (U - V \cos 2 ξ) Γ_{u} = 0 .

(3.10)

The introduction of the parameter $ω^{2} ∕ 4$ can be removed by multiplying both sides of the equation by $4 ∕ ω^{2}$ :

\frac{d^{2} Γ_{u}}{d ξ^{2}} + \frac{8 z e}{m ω^{2} r_{0}^{2}} (U - V \cos 2 ξ) Γ_{u} = 0,

(3.11)

A final simplification of Eq ( 3.11 ) by grouping the parameters into two definitions:

a_{u} = a_{x} = - a_{y} = \frac{8 z e U}{m ω^{2} r_{0}^{2}}, q_{u} = q_{x} = - q_{y} = \frac{4 z e V}{m ω^{2} r_{0}^{2}} .

(3.12)

The current description of the motion equation can be shown as:

\frac{d^{2} Γ_{u}}{d ξ^{2}} + (a_{u} - 2 q_{u} \cos 2 ξ) Γ_{u} = 0

(3.13)

This equation is known in the literature as the Mathieu equation, which was introduced in 1868 by mathematician Emile Mathieu to solve problems related to vibrating elliptic membranes.

Figure 3.9: The stability diagram for the Mathieu equation considering

x

and

y

coordinate directions. Four stability areas are shown in circles. The most common area that is used in quadrupole mass spectrometry is (A).

In a quadrupole the distance between the rods ( $r_{0}$ ) is constant due to construction and the angular frequency ( $ω$ ) is kept constant during operation. Therefore, $U$ and $V$ are the two variables of the analysis. The stability diagram, which Eq. ( 3.13 ) generates, can be seen in Figure 3.9 with the $x$ and $y$ axes being $q_{u}$ and $a_{u}$ . In the figure, there are two distinct areas that can be observed. These coloured areas represent the stability of the travelling ion in the $x$ or $y$ axis. The ion will traverse the quadrupole safely only if the parameters of motion lie in the joint area of both $x$ and $y$ area (i.e., the green area). Rearranging the parameters in Eq. ( 3.12 ) gives:

U = a_{u} \frac{m}{z} \frac{ω^{2} r_{0}^{2}}{8 e} V = q_{u} \frac{m}{z} \frac{ω^{2} r_{0}^{2}}{4 e}

(3.14)

In operation, the ( $ω^{2} r_{0}^{2} ∕ e$ ) part of both $U$ and $V$ is constant, making $m ∕ z$ the only variable. By increasing the $m ∕ z$ value linearly (i.e., switching from one mass to another) the $a_{u}$ and $q_{u}$ will also experience a change in values. This change will cause the triangular area seen in Figure 3.9 , making every mass an individual area in the $V$ - $U$ space. This change in the stability area can be seen in Figure 3.10 .

Figure 3.10: Stability diagrams plotted in RF-DC space, showing a straight scan line through the origin.

Decreasing the slope of the mass scan line allows, the scan line to pass through a major area of the stability diagram. This, turn, widens the mass peak. The outcome of this characteristic shape of the stability diagram is that, as the resolution is reduced (making the peak wider), the position of the leading edge of the mass peak moves to lower apparent mass three times more quickly than the trailing edge of the mass peak moving to a higher apparent mass. This moves the mass peak centre to a lower apparent mass. Therefore, instrument calibration is necessary.

3.3 Conclusion

This chapter establishes the principal experimental parts used throughout the thesis. The transmitter used in the experimental analysis exploits the chemical properties of highly volatile chemicals (VOCs) to generate the chemical signals and the propulsion is generated by $N_{2}$ tank. A descriptive section was dedicated to the underlying theory of QMA along with the membrane inlet technology. As mentioned before, the mass spectrometer (MS) has the ability to detect multiple chemicals simultaneously making MS a valuable receiver for use in molecular communication experimental analysis.

In chapter 4, the mathematical modelling of molecular communications on the macro-scale is described for different environments.

3 Experimental Setup