ATS/Note/2012-041 TECH

2012-05-04

christina.yin.vallgren@cern.ch

Electrical Network Analysis for Vacuum Profile of MedAustron

Christina Yin Vallgren, Paolo Chiggiato, Jose Antonio Ferreira Somoza /TE-VSC

Keywords: MedAustron, Electrical Network Analysis, vacuum,pump-down, MolFlow+

Summary

MedAustron is a synchrotron based hadron therapy facility for cancer treatment cur-rently under construction in Wiener Neustadt, Austria, 40 km south-west of Vienna. Ionbeams of H+

3 and C4+ are generated in gaseous plasma (H2 or CO2) in an Electron Cy-clotron Resonance (ECR) ion source and extracted at a kinetic energy of 8 keV/u andtransformed by the Low-Energy Beam Transfer line (LEBT). The beam of ions is thenaccelerated up to 400 keV/u in the Linear accelerator (LINAC) and transferred via theMedium Energy Beam Transfer line (MEBT) to the synchrotron. After extraction, thebeam is transferred to the treatment rooms for cancer therapy use. The quality of the dosedelivered to the patient for cancer treatment is ultimately determined by the performanceof the beam delivery chain. The performance of beam delivery chain are accordingly in-fluenced by many factors in the entire accelerator chain and the vacuum performance isone of key factors steering the beam quality, especially in the synchrotron. In this re-port, a summary of the simulations done for the entire MedAustron accelerator complexis presented by using so-called Electrical Network Analysis (ENA).

1 Introduction

1.1 Overview of MedAustron

MedAustron is a synchrotron based hadron therapy facility currently under constructionin Wiener Neustadt, Austria, 40 km south-west of Vienna. The decision or constructionwas taken in 2004 by the County of Lower Austria in partnership with the Austrian federalgovernment. With a capacity of 24,000 fractions per year, between 1200 and 1500 patientsper year (depending on fractionation scheme) can be treated with protons and carbon ions[1].

Figure 1 shows the layout of the MedAustron accelerator complex with the transportlines to the irradiation rooms. At the beginning of the injection chain, ions (H+

3 or C4+) are

1

generated in gaseous plasma (H2 or CO2) in an Electron Cyclotron Resonance (ECR) ionsource and extracted at a kinetic energy of 8 keV/u and transformed by the Low-EnergyBeam Transfer line (LEBT). The beam of ions is then accelerated up to 400 keV/u in theLinear accelerator (LINAC), as shown in Fig. 2 and transferred via the Medium EnergyBeam Transfer line (MEBT) to the synchrotron.

After extraction, the beam is transferred to the treatment rooms. Irradiation of thepatient will be performed in three different treatment rooms as indicated in Fig. 1: (1)IR2: Fixed horizontal + fixed vertical beam line (protons and carbon ions). (2) IR3:Fixed horizontal beam line (protons and carbon ions). (3) IR4: Gantry (protons only).In addition, one irradiation room (IR1) is dedicated to non-clinical research.

Figure 1: Layout of MedAustron accelerator chain, including source, LEBT, LINAC,MEBT, synchrotron, extraction line,and transport lines to the treatment rooms[1].

Page 6 of 11 The European Physical Journal Plus

Fig. 5. Injector with Linac, LEBT line and ECR sources.

The major part of the acceleration is performed in the subsequent drift tube linac which is a sequence of 55 drifttubes in the form of an IH structure, with integrated quadrupole triplets for focusing the beam. The RF power forthe RFQ and the IH-structure is provided by three tetrodes operating at 216.8MHz.

3.2.4 Medium-energy beam transfer line

At the end of the injection chain, before the particles are transferred via the MEBT line to the synchrotron, the beamis focused onto a thin 100μg/cm2 carbon foil in order to strip the remaining electrons of the hydrogen and carbon ions.By doing this, the charge-to-mass ratios of the ions are changed to q/m = 1 for protons and q/m = 1/2 for carbonions, which also changes the magnetic rigidities. In the subsequent debuncher a rotation of the beam in longitudinalphase space is performed in order to reduce the energy spread for injection into the synchrotron.

Two special dipoles at the end of the medium energy transfer line act as magnetic injection septa. Their specialconstruction leaves enough space on one side for the adjacent synchrotron vacuum chamber. The magnetic septa arefollowed by a 0.1mm thin electrostatic septum, as the last element of the medium energy transfer line.

3.3 Synchrotron

3.3.1 Overview

Beam transfer from the ion sources to the RFQ/IH linac is activated by a fast electrostatic deflector in the LEBT line,for a nominal pulse duration of about 30μs. The resulting beam pulse is accelerated by the RFQ/IH linac and thenpassed on to the synchrotron where it is injected using a multi-turn injection system. Due to the nature of multi-turninjection a large fraction of the incoming particles are lost, nevertheless the equivalent of about four to five fullyeffective turns are injected. Before starting the RF-acceleration process, the continuous beam has to be captured andbunched by the ring RF-system. At the end of the 50ms long capturing process the beam is collected in one stationarybucket resulting in a particle distribution over about half of the circumference.

Subsequently the synchrotron accelerates the particles to the desired extraction energy, where the RF is switchedoff allowing the beam to debunch before it is extracted by the slow resonance extraction system.

3.3.2 Magnet lattice

The schematic layout of the synchrotron, with injection and extraction elements is shown in fig. 6.The synchrotron lattice has a super-symmetry of two, with a mirror symmetry within each of the two super-periods.

The linear lattice is of the split-FODO type with 16 dipoles, 16 focusing quadrupoles in two circuits and 8 defocusing

Figure 2: Injector with Linac, LEBT line and ECR sources [1].

2

1.2 The aim of this project

The quality of the dose delivered to the patient for cancer treatment is ultimately deter-mined by the performance of the beam delivery chain. The performance of beam deliverychain is accordingly influenced by many factors in the entire accelerator chain and thevacuum performance is one of key factors steering the beam quality, especially in thesynchrotron. The total length of the vacuum chambers in the MedAustron acceleratorcomplex is 363.3 m. The aim of this project is to simulate the entire MedAustron accel-erator vacuum profile by using so-called Electrical Network Analysis (ENA) in order tobe able to:

• monitor pressure at any specific part of accelerator with or without beam.

• monitor pressure with different leakages and different extract gas sources at anyspecific part of accelerator.

• monitor pressure after modification in design, e.g. adding more pumps, movingelements, etc.

• obtain pump-down curves (pressure variation as a function of time) at any specificpart of accelerator in case of intervention and maintenance.

2 Methods

2.1 Calculations of vacuum systems

A schematic diagram of an elementary idealized vacuum system is considered, as shown inFig. 3, in order to derive basic design equations of vacuum calculations [2]. A chamber ofvolume V at a uniform pressure p is connected to a pump set through a component (e.g.stainless steel tube) of conductance C. A gas flow Qin into the system can be caused by apurposely fitted gas inlet, by leaks or permeation through seals or chamber walls. It canalso be caused by gas sources physically within the system like outgassing from materials.

If we apply the ideal gas law with the assumption of constant gas temperature, a fluxbalance yields the basic differential equation:

Vdp

dt= Qin − Qout (1)

where Qout is the flow the pump system removes from the chamber and this can becalculated as:

Qout = p · Seff (2)

where Seff is the effective pumping speed and is determined by:

Seff =1

1S0

+ 1C

(3)

The simplest assumptions, although not necessarily realistic, are pressure- and time-independent pumping speed and injection flow. This yields a simple expression for thepressure as a function of time in the vacuum system:

p(t) = (p0 −Qin

Seff

) · exp(−Seff

Vt) +

Qin

Seff

(4)

3

in

out

𝑝0 𝑆0

𝐶

𝑆𝑒𝑒𝑒

𝑉 𝑝

Figure 3: Schematic diagram of a basic vacuum system. V : volume of vacuum chamber;p: pressure in the vacuum chamber; Qin: gas flow into the system; Qout: flowthe pump system removes from the chamber; S0: pumping speed of the pumpsystem; p0: pressure at the pumps inlet; Seff : effective pumping speed and C:conductance.

Solving Eq. 4 for t yields the pump-down time from p0 to p:

t =V

Seff

· Ln(p0 − Qin

Seff

p− Qin

Seff

) (5)

Conditions such as constant pumping speed or time-independent injection flow arehardly ever met in real systems where transitions from viscous to molecular regime areregularly experienced. Therefore these solutions can provide approximations for real sys-tems; but for sufficiently accurate predictions of the vacuum performance, calculationshave to be based on realistic data of pumps, conductances, and gas flows.

In principle there are two ways to include real data in the design calculations:

1. Model all relevant data such as pumping speeds and flows in an analytical form andsolve the basic equations analytically or

2. Use the data and solve the equations by numerical methods.

Analytical methods yield closed-form solutions which provide insight into the dependen-cies of different parameters. However, analytical solutions are only available for simpleapproximations of, for example, pumping down curves. Numerical methods on the otherhand are very powerful in including the most complex dependencies and system structures.

Since the entire MedAustron vacuum system is very complex, the analytical method isunfortunately not possible. In this work, we use the open source for electrical networkcalculation software (LTSpice) to calculate the vacuum profile of the entire MedAustronaccelerator chain. In the following section, the numerical method used for the MedAustronvacuum system is discussed.

4

2.2 Electrical network analysis for vacuum calculations

It has long been known that an analogy exists between vacuum systems and electricalnetworks. For Electrical Network Analysis (ENA), useful software tools are available.This suggests that we could apply these tools to vacuum system calculations. The relationbetween vacuum metrics and electrical metrics can be found by comparing terms in similarequations, as shown in Table 1.

Electrical metrics Vacuum metrics

Ohm’s law The gas flow

I = G · V Q = C · pCharging/discharging The change ofan electrical capacitor gas content in

of capacitance C a volume V

I = C · dVdt

Q = V · dpdt

I: electrical current [A] Q: flow [Torr l/s]G: electrical conductance [1/Ω] C: conductance of element [l/s]

V: voltage drop [V] p: pressure [Torr]C: capacitance [F] V: volume [l]

Table 1: Relation between electrical metrics and vacuum metrics.

Figure 4 shows a simple example on how to translate a gas flow in a tube into anelectrical network by using commercial simulation software, e.g. LTSpice.

Connection: Electrical Network – Vacuum System?

2012-03-30 TE-VSC seminar MedAustron Vacuum Profile – C. Yin Vallgren 1

Electrical network: By using LTSPICE

Vacuum system: Gas flow into a tube

I: electrical current [A] G: electrical conductance [1 / Ω] V: voltage drop [V] C: capacitance [F]

Q: flow [Torr l / s] C: conductance of element [l/s] p: pressure [Torr] V: volume [l]

Figure 4: Connection between an electrical network and a vacuum system.

2.3 Procedure to calculate the MedAustron vacuum system

In this section, a general procedure on how to simulate a complex vacuum system, suchas the entire MedAustron accelerator chain, by using ENA is presented. The followingsteps are performed:

5

• Step 1: Divide the complete MedAustron vacuum system into many small parts andidentify each element.

Since the MedAustron accelerator is a very complex system and contains manydifferent elements, as shown in Fig. 5, a simplification should be made to easierthe simulation. As seen in Fig. 5, normal chambers, bending chambers in dipole,chambers containing ceramic, vacuum diagnostics tanks with different geometries,as well as vacuum chamber with NEG pump and many other types of vacuumchambers are located in different places in the accelerator chain. The first step ofthis work is to divide the whole accelerator chain into several small parts and thenidentify all the elements in each part.

1

Bending chamber in dipole: Vacuum Diagnostics Tank:

Vacuum chamber with NEG pump:

Chamber with ceramic:

Figure 5: Layout of vacuum chambers in the MedAustron accelerator complex includingdifferent elements, such as chamber with ceramic, bending chamber in dipole,vacuum chamber with NEG pump and different types of vacuum diagnosticstank.

• Step 2: Calculate vacuum parameters for each element.

Once all the elements are identified, we need to calculate vacuum parameters for eachelement, such as conductance, volume and outgassing rate, as already explained inSection. 2.1. To be more pedagogical in explaining how to model a vacuum system byusing ENA, a not so simple example, such as a vacuum diagnostics tank is consideredhere. As seen in Fig. 6, the vacuum diagnostics tank used as example contains onemain tank, one DN160 tube for a turbo molecular pump (TMP) of HiPace 700 l/s,one DN160 tube for beam diagnostics instrument, 6 rectangular chambers for beamdiagnostics instruments and 3 DN40 flanges for installation of gauges. For each ofthe above mentioned elements, the vacuum parameters, as conductance, volume andoutgassing from the surface, are needed to be calculated.

Expression for conductance (Cm) is usefully formulated in term of transmission prob-ability (τ), as shown in Eq. 6. The conductance of a duct is given by the entrance

6

aperture conductance multiplied by the transmission probability. For a simple ge-ometry, τ can easily be found from the general transmission probability tables.However, for complex geometries, Molflow+ [3] by using Monte Carlo simulationcan be applied for the calculations of τ .

Cm = τCa (6)

where Ca is the molecular flow conductance of a thin aperture, which is directlyrelated to the rate of impingement of molecules over the aperture area A.

Ca = A

√R0T

2πMm

(7)

where R0 is the universal gas constant, T is the thermodynamic temperature, andMm is the molar mass (e.g. 0.002 kg/mole for hydrogen).

Outgassing from the surface can be calculated by using measured values:

Q[Torr · l

s · cm2] × surface area [cm2] (8)

Vacuum Diagnostics Tank: 1. Main tank 2. 1xDN160 for

TMP HiPace 700l/s 3. 1xDN160 for

Beam Diagnostics (BD) 4. 6x rectangular for BD 5. 3xDN40 for gauges

TMP: HiPace 700l/s

Degassing rate (H2O) of profile grid monitor after 100h pumping (the number is just an example)

Figure 6: Example of how to model a vacuum diagnostics tank in LTSpice. Left: Thecorresponding electrical network for the different elements in the vacuum diag-nostics tank in LTSpice. Right: The layout of vacuum diagnostics tank, con-taining one main tank, one DN160 tube for a turbo molecular pump (TMP) ofHiPace 700 l/s, one DN160 tube for beam diagnostics instrument, 6 rectangularchambers for beam diagnostics instruments and 3 DN40 flanges for installationof gauges.

• Step 3: Model element in LTSpice program.

7

Once all the elements in a considered vacuum system receive vacuum parameters, wecan then apply the relation between electrical metrics and vacuum metrics and thenuse the LTSpice to simulate the pressure profile. Figure 6 shows an example of thecorresponding electrical network for the different elements in a vacuum diagnosticstank used for MedAustron.

• Step 4: Assembly all elements into one single network.

In Fig. 7, the complete electrical network for the MedAustron accelerator complexfrom the source to the entrance of the synchrotron is shown.

source LEBT LI

MEBT

MR:sector0 1 2 3

4 5 6 7

8 9 10 11

12 13 14 15 0

source LEBT

LI MEBT

MR

1 2

3 4

6

7 8

9

5

10 11

12

13

14

15

0

Figure 7: Assembly all elements of MedAustron accelerator chain into one single network.

3 Results and discussion

In this section, the main results of pressure profile simulated by ENA are presented.Before starting the simulation, the operation pressures from the ion sources have to bedefined. For MedAustron, the ion sources for the production of hydrogen and carbonions are of the ECR type and the detailed description of the gas injection system of theMedAustron ion sources can be found in [4]. Since the ion source used for MedAustronis a well-established and tested system, the operation pressures applied in our pressureprofile calculations are taken directly from the technical report provided by MedAustron[5]. At the present stage, only gas flows and pressures for hydrogen operation have beenmeasured. Therefore the pressure profile simulation is only considered during hydrogenion operation. At a later stage, the pressure profile will also be simulated for the carbonion case. In Table 3 of [5], the pressure measured at the extraction of the ion source islisted. For our calculations, we took 1.7×10−6 mbar as operation pressure for the hydrogenion beam and then assume an upper limit for H2O has a pressure of 4-5 times higher than

8

the measured H2 pressure. For this reason, the H2O pressure at the extraction of the ionsource is assumed to be 7.5×10−6 mbar in all the simulations shown in this work .

3.1 Static pressure profile of H2O after 100 hours pumping

In Fig. 8, pressure profile of H2O after 100 hours pumping at different positions of theMedAustron accelerator complex is shown. The pressure simulation is based on an oper-ational beam with hydrogen ions of 7.5×10−6 mbar at the extraction of the ion source.Two cases have been considered: (1) no outgassing from beam diagnostics instruments istaken into account. (2) outgassing from the Harp Profile Monitor grid (PGX) installed ondifferent vacuum diagnostics tanks in several places of the MEBT, indicated in the ringas seen in Fig. 8 is taken into account. Since only the outgassing from the PGX installedin the MEBT is taken into account in our simulations currently due to the lack of themeasurement data of the other diagnostics instruments used in the accelerator complex,the pressure is expected to be slightly higher in reality. Figure 9 shows the pressurevariation of H2O after 100 hours pumping at different section positions in the Main Ring(MR). As seen in Fig. 9, the crossings between the MR and MEBT contribute highestpressure while the average of the pressure in the rest of the MR stays in 10−10 mbar range.As already mentioned earlier, all the simulations are only considered without taking anyoutgassing from the other beam diagnostics instruments apart from the PGX installed inthe MEBT. It is important to note that the results from the simulations shown in thiswork only provide a insight of the lowest possibility in pressure at present stage.

PGX: Harp Profile Monitor grid installed in MEBT (Outgassing rate: 2.25e-5 mbar l/s after 100 h pumping)

After source: Bare wall: 7.46e-6 mbar w PGX: 7.46e-6 mbar

After S1: Bare wall: 1.4e-7 mbar w PGX: 1.4e-7 mbar

After LE: Bare wall: 2.80e-7 mbar w PGX: 2.80e-7 mbar

After LI: Bare wall: 4.0-9 mbar w PGX: 8.6e-9 mbar

Crossing ME-MR S6-S7: Bare wall: 5.41e-10 mbar w PGX: 1.04e-9 mbar

Crossing ME-MR S0: Bare wall: 2.89e-10 mbar w PGX: 5.6e-9 mbar

MR S11 – S 12: Bare wall: 1.71e-10 mbar w PGX: 1.71e-10 mbar

Figure 8: Pressure profile of H2O after 100 hours pumping with/without considering out-gassing from the Harp Profile Monitor grid (PGX) installed in the MEBT. Theoutgassing rate of the PGX is measured to be 2.25×10−5 mbar·l/s after 100hours pumping, taken from the measurements reported in [6].

9

1.00E-10

1.00E-09

1.00E-08

Pres

sure

[mba

r]

Sector Position in Main Ring

Pressure profile of H2O in Main Ring after 100h pumping

Figure 9: Pressure profile of H2O in the Main Ring after 100 hours pumping consideringoutgassing from the Harp Profile Monitor grid (PGX) installed in the MEBT.

3.2 Pump down curve (MEBT)

In Section 3.1, the pressure profile of H2O after 100 hours pumping is simulated by assum-ing constant pumping speed and constant gas flow in molecular regime. In real systemsthese conditions are certainly not fulfilled for the entire pressure range. Especially, aftereach intervention, it is important to know how the pressure varies with time and how longit takes to pump down a vented system, i.e. to obtain pressure versus time curve - pumpdown curve.

To demonstrate how to calculate a pump down curve after one intervention by usingENA, we chose a sector in the MEBT with two gate valves located at both ends of thesector, as shown in Fig. 10. In Fig. 11, a detailed layout and a list of the different typesof vacuum diagnostics tanks located in the investigated section are shown as well as thepumping groups with respective working pressure ranges are listed.

The water vapor outgassing rate of stainless steel surface as a function of time, is givenby:

Q =2 · 10−9

t· surface area (9)

where t is the pumping time in hour [h].As listed in the table of Fig. 11, the different pumps work at different pressure ranges.

That means an ion pump or a NEG pump will not be able to be operational until thepressure reaches 10−5 mbar. Figure 13 shows the pump-down curve for the investigatedsector with pumping down starting from atmosphere level. As seen in Fig. 13, the gaugeslocated at VDT-H and VDT-E show a high pressure at a level of 10−4 mbar still afterabout 24 hours of pumping. This high pressure is caused by the fact that ion pumpslocated at the same tanks cannot be operational due to the interlock. Two suggestionswould be helpful to reach a reasonable low pressure in this part after one intervention:

1. add an extra turbo-molecular pump (TMP) in-between the VDT-H and the VDT-E

10

to offer more pumping speed in the range of 1×10−4 - 1×10−5 mbar. A mobile turbomolecular pump could be an alternative.

2. improve the material of the PGXs installed at the VDT-H and the VDT-E andmake them more vacuum compatible so that the outgassing in that range decreases,a factor 10 would help significantly.

VVG-R: Gate valve with viton gate sealing DN100

VVG-H: Gate valve with all-metal gate sealing DN100

Figure 10: Layout of the chosen sector for the pump down curve calculation by usingENA.

3.3 NEG pump activation

The sorption characteristics of a NEG pump are generally represented by the sorptionspeed (pumping speed of the NEG) as a function of the sorbed quantity [2]. This quantitycorresponds to the amount of gas adsorbed onto the surface if there is practically no bulkdiffusion, which occurs when NEGs work at room temperature (except for H2). The speedof a NEG can decrease during time, depending on working temperature and pressureconditions. It decreases more rapidly if sorption occurs at room temperature since bulkdiffusion is not promoted and the capacity is basically surface-limited. Therefore it isinteresting to investigate how long a NEG pump works under certain condition until thecapacity of the NEG pump reaches its saturation. When a NEG pump reaches its capacitysaturation, a heating at a temperature of higher than 250C for more than 24 hours isneeded to re-activate adsorption of the material.

For MedAustron, CapaciTorr D 2000 MK 5 pump provided by SAESgetters is used andthis NEG pump is the highest performing pump of the CapaciTorr series. With the useof a special highly porous sintered material, St 172, a Zr-V-Fe getter meterial, the pumpis able to reach 2000 l/s pumping speed for hydrogen in UHV conditions and 1000 l/s forwater. The details about the pump can be found in [7]. As explained earlier, for all thegetter pumps, the pumping speed decreases with increasing of absorbed quantity on thematerial surface.

11

Pumping group

Pump name

Pump speed for H2O (l/s)

Pressure range (mbar)

VPF Scroll Vac

SC15D 4 atm-10-2

HiPace 700 685 10-3 – 10-5

VDT-H /VDT-E (VPI-B)

VacIon plus 300

240 < 10-5

VDT-I (VPI-A)

VacIon plus 500

410 < 10-5

VPN CapaciTorr 2000 MK5

1000 < 10-5

Gate valve

VDT-H

Gate valve

VDT-E

VPF

VDT-I

VPN

PGX: Harp Profile Monitor grid (High outgassing rate) are installed at pumping group VDT-H and VDT-E. Pressure gauges are installed at each pumping group.

Figure 11: Details of the chosen sector for the pump down curve calculation by usingENA.

VDT-H with VacIon Plus 300

VDT-E with VacIon Plus 300 VPF with

Scroll Vac SC 15D HiPace 700

VDT-E with VacIon Plus 300 VPN with CapaciTorr

2000 MK5

Figure 12: Details of the chosen sector for the pump down curve calculation by usingENA.

12

10−5

100

105

10−8

10−6

10−4

10−2

100

102

Time (s)

Pre

ssur

e m

easu

red

at d

iff g

auge

s in

stal

led

at p

ump−

grou

p (m

bar) Pressure as a function of time

Gauge at VDT−HGauge at VDT−EGauge at VDT−IGauge at VVG−H

Figure 13: Pressure [mbar] measured at different gauges installed at pump-groups versustime [s].

In order to calculate the ’time-to-saturation’ of a NEG pump, we first need to calculatethe pumping speed of the NEG as a function of time. In the manual provided by SAESgetter [7], only the experimental data of the pumping speed as a function of sorbed quan-tity is given. By fitting the experimental data, we can easily get an analytical expressionof S = S(Q). The sorbed quantity on the NEG pump can be caused by the outgassingfrom the surrounding materials and the outgassing from the beam diagnostics instrumentsclose by, such as PGX and we can simply get S = S(Q(t)) with Q(t) =

∫ t

0qH2O · dt, where

qH2O is assumed to be the outgassing rate measured on the PGX and the outgassing fromthe stainless steel walls. The detailed calculation is not provided here. In this report, weonly present how the pressure changes at the first crossing between the MEBT and theMR (sector 6 - 7) influenced by sorbed quantity on CapaciTorr NEG pump, see Fig. 14.

As a result, the pressure increases from 1.9×10−9 mbar to 3.4×10−9 mbar after 55 hoursoperation with beam and then stays at 3.47×10−9 mbar level. As conclusion, the pressureat the crossing of the MEBT and the MR does not seem to be so critical and the activationof the NEG pump will only be used to recover from venting to atmosphere.

4 Conclusions and future plans

The simulation of the pressure profile for the MedAustron accelerator complex shows thatthe Electrical Network Analysis (ENA) is indeed a very powerful tool to simulate vacuumprofile in a very complex system when the analytical method is no longer applicable. Sincethe electrical network of the MedAustron accelerator complex is now completed, all thefurther modifications and improvements are easy to add. In order to complete the entiresimulation of the MedAustron vacuum profile and to be able to calculate the pressure asprecisely as possible, all the experimental outgassing rate of all beam diagnostics instru-ments and other items with expected high outgassing rate is needed to be carried out at

13

10−3

10−2

10−1

100

101

100

101

102

103

104

Sorbed Quantity (Torr.l)

Pum

ping

Spe

ed (

l/s)

CapaciTorr D 2000 MK5, S vs Q

(a)

2 2.2 2.4 2.6 2.8 3 3.2 3.4 3.6 3.8

x 105

1.8

2

2.2

2.4

2.6

2.8

3

3.2

3.4x 10

−9

Pre

ssur

e at

cro

ssin

g [m

bar]

Time [s]

Pressure at crossing as a function of time

(b)

Figure 14: Calculation of the capacity of the NEG pump used at the crossing of the MEBTand the MR (sector 6 - 7). (a): Pumping speed of CapaciTorr D2000 MK5[l/s] as a function absorbed quantity on surface by fitting the experimentaldata provided in [7] with two equations. (b): Pressure [mbar] measured at thecrossing of the MEBT and the MR (sector 6 - 7) as a function of time [s].

the lab. Furthermore, it is very important to simulate the detailed pump-down curves forthe suspected critical parts in the accelerator in order to know how many mobile pumpinggroups with specified pumping speed are needed during interventions.

References

[1] M. Benedikt and A. Wrulich, “Medautstron - project overview and status,” THEEUROPEAN PHYSICAL JOURNAL PLUS, vol. 126, p. 69, 2011.

[2] J. M. Lafferty, Foundations of Vacuum Science and Technology. A Wiley-Intersciencepublication, 1997.

[3] R. Kersevan and J.-L. Pons, “Introduction to molflow+,” American Vacuum Society,vol. 27, pp. 1017–1023, 2009.

[4] L. Penescu, “General description of the medaustron ion source. gas system. safety,”ebg MedAustron Technical Note, 2011.

[5] L. Penescu, “Supernanogan ion source - operational data,” ebg MedAustron TechnicalNote, 2012.

[6] D. P. Holzer, S. Meunier, and I. Wevers, “Pumpdown measurement of various compo-nents for the medaustron wire scanner,” TE-VSC Technical Note, 2011.

[7] SAEAgetters, “Capacitorr pumps mk5 series,” 2001.

14