|
Linacs have many applications: they generate [[X-ray]]s and high energy electrons for medicinal purposes in [[radiation therapy]], serve as particle injectors for higher-energy accelerators, and are used directly to achieve the highest kinetic energy for light particles (electrons and positrons) for [[particle physics]].
The design of a linac depends on the type of particle that is being accelerated: [[electron]]s, [[proton]]s or [[ion]]sions. Linacs range in size from a [[cathode -ray tube]] (which is a type of linac) to the {{convert|3.2|km|mi|adj=mid|-long}} linac at the [[SLAC National Accelerator Laboratory]] in [[Menlo Park, California]].
==History==
[[File:Wideroe linac en.svg|thumb|300px|Wideroe's linac concept. The voltage from an RF source is connected to a series of tubes which shield the particle between gaps.]]
[[File:CERN Linac1.jpg|thumb|Alvarez type linac]]
In 1924, [[Gustaf Ising|Gustav Ising]] published the first description of a linear particle accelerator using a series of accelerating gaps. Particles would proceed down a series of tubes. At a regular frequency, an accelerating voltage would be applied across each gap. As the particles gained speed while the frequency remained constant, the gaps would be spaced farther and farther apart, in order to ensure the particle would see a voltage applied as it reached each gap. Ising never successfully implemented this design.<ref name="heibron">{{cite book |last1=Heilbron |first1=J.L. |last2=Seidel |first2=Robert W. |title=Lawrence and His Laboratory: A History of the Lawrence Berkeley Laboratory, Volume I |date=1989 |publisher=University of California Press |___location=Berkeley, CA |url=http://ark.cdlib.org/ark:/13030/ft5s200764/ |access-date=2 February 2022}}</ref>
[[Rolf Wideroe]] discovered Ising's paper in 1927, and as part of his PhD thesis, he built an 88-inch long, two gap version of the device. Where Ising had proposed a spark gap as the voltage source, Wideroe used a 25kV [[vacuum tube]] oscillator. He successfully demonstrated that he had accelerated sodium and potassium ions to an energy of 50 thousand,000 [[electron volt]]s (50 keV), twice the energy they would have received if accelerated only once by the tube. By successfully accelerating a particle multiple times using the same voltage source, Wideroe demonstrated the utility of [[radio frequency]] (RF) acceleration.<ref>{{cite book |last1=Conte |first1=Mario |last2=MacKay |first2=William |title=An introduction to the physics of particle accelerators |date=2008 |publisher=World Scientific |___location=Hackensack, N.J. |isbn=9789812779601 |edition=2nd}}</ref>
This type of linac was limited by the voltage sources that were available at the time, and it was not until after [[World War II]] that [[Luis Walter Alvarez|Luis Alvarez]] was able to use newly developed high frequency oscillators to design the first resonant cavity drift tube linac. An Alvarez linac differs from the Wideroe type in that the RF power is applied to the entire [[Resonator#Cavity resonators|resonant chamber]] through which the particle travels, and the central tubes are only used to shield the particles during the decelerating portion of the oscillator's phase. Using this approach to acceleration meant that Alvarez's first linac was able to achieve proton energies of 31.5 MeV in 1947, the highest that had ever been reached at the time.<ref>{{cite web |title=Alvarez proton linear accelerator |url=https://www.si.edu/es/object/alvarez-proton-linear-accelerator%3Anmah_700150 |website=Smithsonian Institution |access-date=3 February 2022 }}</ref>
The initial Alvarez type linacs had no strong mechanism for keeping the beam focused, and were limited in length and energy as a result. The development of the [[strong focusing]] principle in the early 1950s led to the installation of focusing [[quadrupole magnet]]s inside the drift tubes, allowing for longer and thus more powerful linacs. Two of the earliest examples of Alvarez linacs with strong focusing magnets were built at [[CERN]] and [[Brookhaven National Laboratory]].<ref>{{cite report |author=Lapostolle, Pierre |date= July 1989 |title= Proton Linear Accelerators: A Theoretical and Historical Introduction |url= https://www.osti.gov/servlets/purl/6038195 |publisher= Los Alamos National Laboratory |docket=LA-11601-MS |access-date= February 4, 2022}}</ref>
In 1947, at about the same time that Alvarez was developing his linac concept for protons, [[W. W. Hansen|William Hansen]] constructed the first travelling-wave electron accelerator at Stanford University.<ref>{{cite journal |author-last=Ginzton |author-first=Edward L. |date= April 1983| title=Early Accelerator Work at Stanford |journal= SLAC Beam Line | pages=2–16 |url=http://atlas.physics.arizona.edu/~shupe/Physics_Courses/Phys_586_S2015_S2016_S2017/Readings_MS/SLAC_Early_History.pdf
}}</ref> Electrons are sufficiently lighter than protons that they achieve speeds close to the [[speed of light]] early in the acceleration process. As a result, "accelerating" electrons increase in energy, but can be treated as having a constant velocity from an accelerator design standpoint. This allowed Hansen to use an accelerating structure consisting of a horizontal [[waveguide]] loaded by a series of discs. The 1947 accelerator had an energy of 6 MeV. Over time, electron acceleration at the [[SLAC National Accelerator Laboratory]] would extend to a size of {{convert|2|mi|km}} and an output energy of 50 GeV.<ref>{{cite book |last1=Neal |first1=R. B. |title=The Stanford Two-Mile Accelerator |chapter=Chap. 5 |publisher=W.A. Benjamin, Inc |year=1968 |___location=New York, New York |page=59 |chapter-url=http://www.slac.stanford.edu/spires/hep/HEPPDF/twomile/Chapters_4_5.pdf |access-date=2010-09-17}}</ref>
As linear accelerators were developed with higher beam currents, using magnetic fields to focus proton and heavy ion beams presented difficulties for the initial stages of the accelerator. Because the [[Lorentz force|magnetic force]] is dependent on the particle velocity, it was desirable to create a type of accelerator which could simultaneously accelerate and focus low-to-mid energy [[hadron]]s.<ref>{{cite journal |last1=Stokes |first1=Richard H. |last2=Wangler |first2=Thomas P. |title=Radiofrequency Quadrupole Accelerators and their Applications |journal=Annual Review of Nuclear and Particle Science |date=1988 |volume=38 |issue=38 |pages=97–118 |doi=10.1146/annurev.ns.38.120188.000525 |bibcode=1988ARNPS..38...97S |url=https://www.annualreviews.org/doi/pdf/10.1146/annurev.ns.38.120188.000525 |access-date=3 February 2022|doi-access=free }}</ref> In 1970, Soviet physicists I. M. Kapchinsky and [[Vladimir Teplyakov]] proposed the [[Radio-frequency quadrupole|radio-frequency quadrupole]] (RFQ)]] type of accelerating structure. RFQs use vanes or rods with precisely designed shapes in a resonant cavity to produce complex electric fields. These fields provide simultaneous acceleration and focusing to injected particle beams.<ref name = "Reiser 2008, p6">{{cite book |title= Theory and design of charged particle beams |last1= Reiser |first1= Martin |edition = 2nd |date= 2008 |publisher= [[Wiley-VCH]] |___location= Weinheim |isbn= 9783527407415 |page=6 |url= https://books.google.com/books?id=eegK9Mqgpi4C}}</ref>
Beginning in the 1960s, scientists at Stanford and elsewhere began to explore the use of [[superconducting radio frequency]] cavities for particle acceleration.<ref>{{cite arXiv | last=Padamsee | first=Hasan | date= April 14, 2020 | title= History of gradient advances in SRF | class=physics.acc-ph | eprint=2004.06720}}</ref> Superconducting cavities made of [[niobium]] alloys allowedallow for much more efficient acceleration, as a substantially higher fraction of the input power could be applied to the beam, rather than lost to heat. Some of the earliest superconducting linacs included the Superconducting Linear Accelerator (for electrons) at Stanford<ref>{{cite report | first=Catherine | last=Westfall | title=The Prehistory of Jefferson Lab's SRF Accelerating Cavities, 1962 to 1985 | date=April 1997 | publisher=[[Thomas Jefferson National Accelerator Facility]] | docket=JLAB-PHY-97-35 | url=https://misportal.jlab.org/ul/publications/view_pub.cfm?pub_id=11132}}</ref> and the [[Argonne Tandem Linear Accelerator System]] (for protons and heavy ions) at [[Argonne National Laboratory]].<ref>{{cite journal |last1=Ostroumov |first1=Peter |last2=Gerigk |first2=Frank |title=Superconducting Hadron Linacs |journal=Reviews of Accelerator Science and Technology |date=January 2013 |volume=06 |pages=171–196 |doi=10.1142/S1793626813300089}}</ref>
==Basic principles of operation==
=== Radiofrequency acceleration ===
When a [[charged particle]] is placed in an [[electromagnetic field]] it experiences a force given by the [[Lorentz force law]] law:
:<math>\vec{F} = q \vec{E} + q \vec{v} \times \vec{B}</math>
(in SI units) where <math>q</math> is the charge on the particle, <math>\vec{E}</math> is the electric field, <math>\vec{v}</math> is the particle velocity, and <math>\vec{B}</math> is the magnetic field. The cross product in the magnetic field term means that static magnetic fields cannot be used for particle acceleration, as the magnetic force acts perpendicularly to the direction of particle motion.<ref name="conte">{{cite book |last1=Conte |first1=Mario |last2=MacKay |first2=William |title=An introduction to the physics of particle accelerators |date=2008 |publisher=World Scientific |___location=Hackensack, N.J. |isbn=9789812779601 |pages=1 |edition=2nd}}</ref>
As [[Electrical breakdown|electrostatic breakdown]] limits the maximum constant voltage which can be applied across a gap to produce an electric field, most accelerators use some form of [[radiofrequency]] (RF) acceleration. In RF acceleration, the particle traverses a series of accelerating regions, driven by a source of voltage in such a way that the particle sees an accelerating field as it crosses each region. In this type of acceleration, particles must necessarily travel in "bunches" corresponding to the portion of the oscillator's cycle where the electric field is pointing in the intended direction of acceleration.<ref name="edwards">{{cite book |last1=Edwards |first1=D. A. |last2=Syphers |first2=M.J. |title=An introduction to the physics of high energy accelerators |date=1993 |publisher=Wiley |___location=New York |isbn=9780471551638}}</ref>
If a single oscillating voltage source is used to drive a series of gaps, those gaps must be placed increasingly far apart as the speed of the particle increases. This is to ensure that the particle "sees" the same phase of the oscillator's cycle as it reaches each gap. As particles asymptotically approach the speed of light, the gap separation becomes constant -: additional applied force increases the energy of the particles, but does not significantly alter their speed.{{r|conte|p=9-12}}
=== Focusing ===
In order to ensure particles do not escape the accelerator, it is necessary to provide some form of focusing to redirect particles moving away from the central trajectory back towards the intended path. With the discovery of [[strong focusing]], [[quadrupole magnets]] are used to actively redirect particles moving away from the reference path. As quadrupole magnets are focusing in one transverse direction and defocusing in the perpendicular direction, it is necessary to use groups of magnets to provide an overall focusing effect in both directions.{{r|conte}}
====Phase stability====
Focusing along the direction of travel, also known as ''phase stability'', is an inherent property of RF acceleration. If the particles in a bunch all reach the accelerating region during the rising phase of the oscillating field, then particles which arrive early will see slightly less voltage than the "reference" particle at the center of the bunch. Those particles will therefore receive slightly less acceleration and eventually fall behind the reference particle. Correspondingly, particles which arrive after the reference particle will receive slightly more acceleration, and will catch up to the reference as a result. This automatic correction occurs at each accelerating gap, so the bunch is refocused along the direction of travel each time it is accelerated.{{r|edwards|pp=30-52}}
==Construction and operation==
A linear particle accelerator consists of the following parts:
*A straight hollow pipe [[vacuum chamber]] which contains the other components. It is evacuated with a [[vacuum pump]] so that the accelerated particles will not collide with air molecules. The length will vary with the application. If the device is used for the production of X-rays for inspection or therapy, then the pipe may be only 0.5 to 1.5 meters long.<ref>{{cite book |last1=Podgorsak |first1=E B |title=Radiation Oncology Physics |date=2005 |publisher=[[International Atomic Energy Agency]] |___location=Vienna |isbn=92-0-107304-6 |page=138 |url=https://www.iaea.org/publications/7086/radiation-oncology-physics |language=en |chapter=Treatment Machines for External Beam Radiotherapy}}</ref> If the device is to be an injector for a [[synchrotron]], it may be about ten meters long.<ref>{{cite conference |url=https://s3.cern.ch/inspire-prod-files-b/bf2698b68b2ac10829c592a2137d3ee2 |title=Linear Accelerator Injectors for Proton Synchrotrons |first=J P |last=Blewett |author= |author-link= |date=11 June 1956 |conference=CERN Symposium on High Energy Accelerators and Pion Physics |conference-url= |editor=Edouard Regenstreif |volume=1 |edition= |book-title= |publisher=CERN |archive-url= |archive-date= |___location=Geneva |pages=159–166 |format= |id= |isbn= |bibcode= |oclc= |doi=10.5170/CERN-1956-025 |access-date= |quote= |language= |page= |at= |trans-title=}}</ref> If the device is used as the primary accelerator for nuclear particle investigations, it may be several thousand meters long.<ref>{{cite web |last1=Weise |first1=Hans |last2=Decking |first2=Winfried |title=The world's longest superconducting linac |url=https://cerncourier.com/a/the-worlds-longest-superconducting-linac/ |website=[[CERN Courier]] |publisher=IOP Publishing |date=10 July 2017}}</ref>
*The particle source ''(S)'' at one end of the chamber which produces the [[charged particle]]s which the machine accelerates. The design of the source depends on the particle that is being accelerated. [[Electron]]s are generated by a [[cold cathode]], a [[hot cathode]], a [[photocathode]], or [[RF antenna ion source|radio frequency (RF) ion sources]]. [[Proton]]s are generated in an [[ion source]], which can have many different designs. If heavier particles are to be accelerated, (e.g., [[uranium]] [[ions]]), a specialized ion source is needed. The source has its own high voltage supply to inject the particles into the beamline.<ref>{{cite arXiv |last=Faircloth |first=D C |author-link= |eprint=2103.13231 |title=Particle Sources |class= physics.acc-ph|date=24 March 2021 }}</ref>
*Extending along the pipe from the source is a series of open-ended cylindrical electrodes ''(C1, C2, C3, C4)'', whose length increases progressively with the distance from the source. The particles from the source pass through these electrodes. The length of each electrode is determined by the frequency and power of the driving power source and the particle to be accelerated, so that the particle passes through each electrode in exactly one-half cycle of the accelerating voltage. The mass of the particle has a large effect on the length of the cylindrical electrodes; for example an electron is considerably lighter than a proton and so will generally require a much smaller section of cylindrical electrodes as it accelerates very quickly.
*A target ''(not shown)'' with which the particles collide, located at the end of the accelerating electrodes. If electrons are accelerated to produce [[X-rays]], then a water -cooled [[tungsten]] target is used. Various target materials are used when [[proton]]sprotons or other nuclei are accelerated, depending upon the specific investigation. Behind the target are various detectors to detect the particles resulting from the collision of the incoming particles with the atoms of the target. Many linacs serve as the initial accelerator stage for larger particle accelerators such as [[synchrotron]]s and [[storage ring]]s, and in this case after leaving the electrodes the accelerated particles enter the next stage of the accelerator.
*An [[electronic oscillator]] and [[amplifier]] ''(G)'' which generates a [[radio frequency]] [[alternating current|AC]] [[voltage]] of high potential (usually thousands of volts) which is applied to the cylindrical electrodes. This is the accelerating voltage which produces the electric field which accelerates the particles. As shown, oppositeOpposite phase voltage is applied to successive electrodes. A high power accelerator will have a separate amplifier to power each electrode, all synchronized to the same frequency.
As shown in the animation, the oscillating voltage applied to alternate cylindrical electrodes has opposite polarity (180° [[out of phase]]), so adjacent electrodes have opposite voltages. This creates an oscillating [[electric field]] ''(E)'' in the gap between each pair of electrodes, which exerts force on the particles when they pass through, imparting energy to them by accelerating them. The particle source injects a group of particles into the first electrode once each cycle of the voltage, when the charge on the electrode is opposite to the charge on the particles. The electrodes are made the correct length so that the accelerating particles take exactly one-half cycle to pass through each electrode. Each time the particle bunch passes through an electrode, the oscillating voltage changes polarity, so when the particles reach the gap between electrodes the electric field is in the correct direction to accelerate them. Therefore, the particles accelerate to a faster speed each time they pass between electrodes; there is little electric field inside the electrodes so the particles travel at a constant speed within each electrode.
The particles are injected at the right time so that the oscillating voltage differential between electrodes is maximum as the particles cross each gap. If the peak voltage applied between the electrodes is <math>V_p</math> [[volt]]svolts, and the charge on each particle is <math>q</math> [[elementary charge]]s, the particle gains an equal increment of energy of <math>qV_p</math> [[electron volt]]s when passing through each gap. Thus the output energy of the particles is
:<math>E = qNV_p</math>
electron volts, where <math>N</math> is the number of accelerating electrodes in the machine.
At speeds near the speed of light, the incremental velocity increase will be small, with the energy appearing as an increase in the mass of the particles. In portions of the accelerator where this occurs, the tubular electrode lengths will be almost constant. Additional magnetic or electrostatic lens elements may be included to ensure that the beam remains in the center of the pipe and its electrodes. Very long accelerators may maintain a precise alignment of their components through the use of servo systems guided by a laser beam.
== Concepts in development ==
=== Induction linear accelerator ===
Induction linear accelerators use the electric field induced by a time-varying magnetic field for acceleration - likeacceleration—like the [[betatron]]. The particle beam passes through a series of ring-shaped [[ferrite core]]s standing one behind the other, which are magnetized by high-current pulses, and in turn each generate an electrical field strength pulse along the axis of the beam direction. Induction linear accelerators are considered for short high current pulses from electrons, but also from heavy ions.<ref>{{Cite web|date=2002-06-25|title=Heavy ions offer a new approach to fusion|url=https://cerncourier.com/a/heavy-ions-offer-a-new-approach-to-fusion/|access-date=2021-01-22|website=CERN Courier|language=en-GB}}</ref> The concept goes back to the work of [[Nicholas Christofilos]].<ref>{{Cite journal|last1=Christofilos|first1=N. C.|last2=Hester|first2=R. E.|last3=Lamb|first3=W. a. S.|last4=Reagan|first4=D. D.|last5=Sherwood|first5=W. A.|last6=Wright|first6=R. E.|date=1964-07-01|title=High Current Linear Induction Accelerator for Electrons|url=https://aip.scitation.org/doi/10.1063/1.1746846|journal=Review of Scientific Instruments|volume=35|issue=7|pages=886–890|doi=10.1063/1.1746846|bibcode=1964RScI...35..886C|issn=0034-6748}}</ref> Its realization is highly dependent on progress in the development of more suitable [[Ferrite (magnet)|Ferriteferrite]] materials. With electrons, pulse currents of up to 5 kiloamps at energies up to 5 MeV and pulse durations in the range of 20 to 300 nanoseconds were achieved.<ref>{{Cite journal|last=Fraas|first=H.|date=1989|title=Kern- und Elementarteilchenphysik. Von G. Musiol, J. Ranft, R. Reif und D. Seeliger, VCH Verlagsgesellschaft Weinheim, 1988, DM 128|url=http://adsabs.harvard.edu/abs/1989PhuZ...20...31F|journal=Physik in unserer Zeit|volume=20|issue=1|pages=31|doi=10.1002/piuz.19890200109|bibcode=1989PhuZ...20...31F|issn=0031-9252}}</ref>
=== Energy Recoveryrecovery LINAClinac ===
In previous electron linear accelerators, the accelerated particles are used only once and then fed into an absorber ''(beam dump)'', in which their residual energy is converted into heat. In an ''Energyenergy Recoveryrecovery Linac''linac (ERL; literally: "Energy recovery linear accelerator"), instead, the accelerated in resonators and, for example, in [[undulator]]s. The electrons used are fed back through the accelerator, out of phase by 180 degrees. They therefore pass through the resonators in the decelerating phase and thus return their remaining energy to the field. The concept is comparable to the hybrid drive of motor vehicles, where the kinetic energy released during braking is made available for the next acceleration by charging a battery.
The [[Brookhaven National Laboratory]] and the [[Helmholtz-Zentrum Berlin]] with the project "bERLinPro" reported on corresponding development work . The Berlin experimental accelerator uses superconducting niobium cavity resonators of the type mentioned above. In 2014, three [[free-electron laser]]s based on ''Energy Recovery LinacsERLs were'' in operation ''worldwide'' : in the [[Thomas Jefferson National Accelerator Facility|Jefferson Lab]] (USAUS), in the [[Budker Institute of Nuclear Physics]] (Russia) and at JAEA (Japan).<ref>{{Cite book|url=https://www.springer.com/gp/book/9783319143934|title=Synchrotron Light Sources and Free-Electron Lasers: Accelerator Physics, Instrumentation and Science Applications|date=2016|publisher=Springer International Publishing|isbn=978-3-319-14393-4|editor-last=Jaeschke|editor-first=Eberhard|language=en|editor-last2=Khan|editor-first2=Shaukat|editor-last3=Schneider|editor-first3=Jochen R.|editor-last4=Hastings|editor-first4=Jerome B.}}</ref> At the [[University of Mainz]] , an ERL called MESA is underexpected constructionto and should (as of 2019) go intobegin operation in 20222024.<ref>https://www.prisma.uni-mainz.de/facilities/mesa-mainz-energy-recovering-superconducting-accelerator/</ref>
=== Compact Linear Collider ===
The concept of the [[Compact Linear Collider]] (CLIC) (original name ''CERN Linear Collider'' , with the same abbreviation) for electrons and positrons provides a traveling wave accelerator for energies of the order of 1 tera-electron volt (TeV).<ref>{{Cite book|last=Raubenheimer|first=T. O.|title=A 3 TeV e+e− linear collider based on CLIC technology|publisher=|year=2000|isbn=92-9083-168-5|___location=Geneva|pages=}}</ref> Instead of the otherwise necessary numerous [[klystron]] amplifiers to generate the acceleration power, a second, parallel electron linear accelerator of lower energy is to be used, which works with superconducting cavities in which standing waves are formed. High-frequency power is extracted from it at regular intervals and transmitted to the main accelerator. In this way, the very high acceleration field strength of 80 MV / m should be achieved.
=== Kielfeld accelerator (plasma accelerator) ===
In cavity resonators, the dielectric strength limits the maximum acceleration that can be achieved within a certain distance. This limit can be circumvented using accelerated waves in plasma to generate the accelerating field in [[Plasma acceleration|Kielfeld accelerators]]: A laser or particle beam excites an oscillation in a [[Plasma (physics)|plasma]], which is associated with very strong electric field strengths. This means that significantly (factors of 100s to 1000s ) more compact linear accelerators can possibly be built. Experiments involving high power lasers in metal vapour plasmas suggest that a beam line length reduction from some tens of metres to a few cm is quite possible.
===Compact Medicalmedical Acceleratorsaccelerators ===
The LIGHT program (Linac for Image-Guided Hadron Therapy) hopes to create a design capable of accelerating protons to 200MeV or so for medical use over a distance of a few tens of metres, by optimising and nesting existing accelerator techniques <ref>{{Cite web| url=http://cdsweb.cern.ch/record/2314160/files/frb1io02.pdf| title =LIGHT: A LINEAR ACCELERATOR FOR PROTON THERAPY}}</ref> The current design (2020) uses the highest practical bunch frequency (currently ~ 3 GHz) for a [[Radio-frequency quadrupole]] (RFQ) stage from injection at 50kVdC to ~5MeV bunches, a Side Coupled Drift Tube Linac (SCDTL) to accelerate from 5Mev to ~ 40MeV and a Cell Coupled Linac (CCL) stage final, taking the output to 200-230MeV. Each stage is optimised to allow close coupling and synchronous operation during the beam energy build-up.
The project aim is to make proton therapy a more accessible mainstream medicine as an alternative to existing radio therapy.
== Modern linear accelerator concepts ==
The higher the frequency of the acceleration voltage selected, the more individual acceleration thrusts per path length a particle of a given speed experiences, and the shorter the accelerator can therefore be overall. That is why accelerator technology developed in the pursuit of higher particle energies, especially towards higher frequencies.
[[File:Linear Accelerator.jpg|thumb|right|250px|Steel casting undergoing x-ray using the linear accelerator at [[Goodwin Steel Castings Ltd]]]]
The linear accelerator could produce higher particle energies than the previous [[electrostatic particle accelerator]]s (the [[Cockcroft-Walton accelerator]] and [[Van de Graaff generator]]) that were in use when it was invented. In these machines, the particles were only accelerated once by the applied voltage, so the particle energy in [[electron volts]] was equal to the accelerating voltage on the machine, which was limited to a few million volts by insulation breakdown. In the linac, the particles are accelerated multiple times by the applied voltage, so the particle energy is not limited by the accelerating voltage.
High power linacs are also being developed for production of electrons at relativistic speeds, required since fast electrons traveling in an arc will lose energy through [[synchrotron radiation]]; this limits the maximum power that can be imparted to electrons in a synchrotron of given size.
| 