Hatano Y., Katsumura Y., Mozumder A. (Eds.) Charged Particle and Photon Interactions with Matter - Recent Advances, Applications, and Interfaces

Подождите немного. Документ загружается.

580 Charged Particle and Photon Interactions with Matter

These are as follows:

1. Production

of a stable radiation beam of micron or submicron size.

2. Suitable

dose rate to assure a fast and accurate dose delivery.

3. Radiation detectors able to monitor, with high efciency, the dose delivered to the samples

and trigger the beam-stop mechanism when the desired dose has been reached.

4. Specially designed cell holder to maintain the sample in a suitable sterile environment

while

allowing cell visualization and irradiation.

5. Image system for sample localization. This will have to be supported by appropriate soft-

ware

for image analysis and coordinate recording.

6. Micropositioning stage to align the samples with the radiation probe with high spatial reso-

lution and reproducibility. The stage speed is also critical in assuring the high throughput

necessary

to investigate low frequency events (i.e., oncogenic transformation).

21.5 Charged-partiCle miCrobeams

21.5.1 MicrobeaM production

The key aspect for any microbeam facility is the ability to spatially limit the irradiation to a specic

sample or part of it. There are two methods to reduce an accelerated beam of charged particles to a

micron

or submicron sized spot: collimation and focusing.

Collimation

is by far the more straightforward way offering a number of advantages such as

beam location, easy extraction into air, and reduction of the particle ux to radiobiologically rel-

evant dose-rate ranges. As the samples need to be placed as close as possible to the collimator end,

this is normally achieved by the optical devices used to image the sample allowing therefore a very

precise beam location. Moreover, the size of the collimators represents a natural way to cut off the

high ux of particles generated by the accelerators making it easy to reduce the uency to ∼100s of

particles per second. Such a rate is ideal for irradiating a large number of samples in a short period

of time while achieving a very high accuracy in the number of particles delivered (see particle

detection & delivery). The main limitation of the collimation approach is particle scattering, which

allows low energy particles to emerge with a wider lateral angle than the primary beam (this effect

is inversely proportional to the radius–length ratio of the collimator). Scattering can be minimized

by optimizing the collimator material, geometry, and beam alignment, all of which require sig-

nicant design and skills. At the Gray Laboratory, several collimator options have been evaluated

with the best being a 1 mm length of thick walled (outside diameter of 225 μm + 10μm polyimide

sheath) fused silica tubing with apertures as small as 1μm in diameter (Folkard etal., 1997a,b).

The collimator was mounted into a gimballing assembly (which allowed a precise beam–collimator

alignment) and capped by a 3μm Mylar lm that acts as a vacuum window. The whole assembly

was motorized in the vertical direction to move the end of the collimator closer to the sample during

the irradiation and back down to allow the next sample positioning. The collimator performance

was assessed in terms of quality of the energy and spatial spread. The rst was done by measuring

the energy spectrum of the emerging particles, resulting in a full width at half-maximum of 47keV,

while the latter was calculated by measuring the distance between particles tracks produced by

irradiating CR-39 track-etch detectors. Using protons, it was possible to conne 90% of the par-

ticles within a 2 μm spot and 96% within 5 μm, while using

, an accuracy of 99% <2μm was

achieved (Folkard etal., 2005). An alternative collimation system is represented by the use of a pair

of aligned microapertures, where the rst one denes the beam size and the second (usually larger)

acts as an antiscatter device stopping the particles scattered by the rst aperture. The distance

between the two apertures and the relative alignment with the beam are critical parameters. Such

a system has been adopted by many laboratories such as the JAERI (Takasaki, Japan) (Kobayashi

etal., 2004), the INFN-LNL (Italy) (Gerardi etal., 2005), and the Columbia University where a pair

Application of Microbeams to the Study of the Biological Effects of Low Dose Irradiation 581

of laser-drilled microapertures (5 and 6μm) separated by 300μm was assembled at the end of the

beam line using two orthogonal micrometers (Randers-Pehrson etal., 2000). Monte Carlo modeling

indicates that 91% of the emerging particles are within the unscattered area (5 μm diameter) with

of the beam diffused throughout an 8

μm

halo.

collimation was the preferred method of the rst modern microbeams, the focusing approach

has rapidly increased in popularity, mainly driven by already existing facilities (previously used

for nuclear microscopy) and the need to investigate the sensitivity of subcellular targets such as

the mitochondria (Tartier etal., 2007). There is little doubt that the narrowest charged-particle

beams (down to a few 10s of nanometers) and high targeting accuracy are achieved in a vacuum by

magnetic or electrostatic focusing devices as they don’t suffer from collimator scattering problems.

However, radiobiological microbeams require the radiation beam to be extracted in air in order

to irradiate living samples in their natural surroundings. As the focused beam passes through the

vacuum window, signicant scattering is introduced, which compromises and limits the achievable

beam spot size. Focused spots of 1μm or less on the samples are, however, achievable with a very

low fraction of scattered particles of reduced energy reaching the targets. Other relevant factors to

consider for microbeam facilities based on focusing devices are a usually high beam current (which

may limit their application for low-dose studies), the nonnegligible costs of the devices, and the

length required to achieve signicant focusing. The latter point is usually a strong constraint that

limits the adoption of focusing lenses for radiobiological microbeams on a wide scale. Finally, a

further complication of focusing-based microbeams is the localization of the radiation probe. While

this is a straightforward operation when using collimators (whose end can be visualized using the

same optics used for the biological sample), localizing the focused radiation spot in 3D requires

extra and more complex steps. Once the ion beam focus has been achieved (this by itself may require

iterative sequences measuring the ion beam size for combinations of the electrostatic lens voltages),

the X–Y beam location is generally determined by scanning a knife edge (or a known shape object)

across the beam prole while monitoring the particle energy-loss. When the knife edge is placed at

the center of the ion beam, it is visualized registering the actual beam position to which to align the

samples. Successful charged-particle microbeams based on focus systems have been developed and

are routinely used in Germany at the PTB (Greif etal., 2006) and at the GSI (Heiss etal., 2004), in

Japan at the NIRS (Imaseki etal., 2007), and at the Columbia University (United States). Brenner

and colleagues (Bigelow etal., 2009) have adopted a compound lens consisting of two electrostatic

quadrupole triplets with “Russian symmetry” to ensure a circular beam spot of 0.8μm for 6 MeV He

ions. The lenses have been realized using four 1 cm diameter ceramic rods, gold plated in bands to

create positive and negative electrodes whose strengths are arranged following the (+A, −B, +B, −A)

pattern for each quadruplet and (+A, −B, +C), (−C, +B, −A) for the triplets with voltage up to 15kV.

Each triplet is 0.3m long and they are placed 1.9m apart.

21.5.2 beaM detection

Another key feature of the modern microbeam facilities is the ability to detect and therefore control

the number of particles delivered to the samples. In order to achieve this, it is necessary to develop a

high-efciency detection system (which will trigger the signal when the preset number of events has

been reached) coupled to a very fast beam deection system to terminate the irradiation. Particle

detection is probably the feature that differs most between the microbeam facilities developed so

far. They basically can be divided into two categories according to whether the detection occurs

before

or after the particle reaches the biological sample.

placing the detector between the vacuum window and the samples, no further constraints

are imposed on the sample holder or the cell environment. Moreover, it is possible to use par-

ticles of energy low enough to stop within the cell. The main issue with this type of approach

is the limited space available and the inevitable detector–beam interaction, which reduces the

quality and accuracy of the exposure. In order to minimize energy loss in the detector, only

582 Charged Particle and Photon Interactions with Matter

thin, transmission-type detectors are appropriate. These abrogate the effects of increased beam

size and reduced targeting accuracy due to scattering. These detectors are generally thin lms

of plastic scintillators whose light ashes generated by the particle traversals are collected by a

photomultiplier and then processed. It is also critical to choose a scintillator with a short decay

time (<100 ns) to avoid light ashes generated by different particles that may overlap resulting in

false signals. Available scintillators are not very efcient, and their reduced thickness leads to

very weak light signals. A 5 MeV α particle loses ∼300 keV in a 5μm plastic scintillator, of which

only ∼4% are converted in 2000–3000 isotropically emitted photons of 3–5eV. It is therefore

necessary to collect most of the emitted photons using a very efcient photomultiplier and placing

them close to the scintillator without interfering with the sample irradiation. This approach was

rst successfully adopted by the Gray Laboratory (Folkard etal., 1997a,b) using a 18 μm BC400

scintillator and an Hamamatsu PM-tube switched into position of the microscope objective (i.e.,

immediately above the scintillator and the samples) during the irradiation. By enclosing the irra-

diation area in a light-proof cage, a detection efciency of >99% was achieved for both protons

and

ions. Slight variation of the above method has been followed at the CENBG. Here,

a custom-designed low pressure transmission gas chamber (3.5 mm thick, lled with 10mbar

isobutene) with 100 nm Si

windows is placed in the beam path upstream of the samples. The

system has been reported to monitor the particle traversals with greater efciency and minimum

interference (Barberet etal., 2005).

The alternative conguration consists of placing the detector behind the sample holder and

therefore monitoring the particles after they have been delivered to the cells. Using this approach,

there is no extra scattering introduced by the detector, and in theory better targeting accuracy

can be reached. Such congurations benet also from less space constraints for the detector, and

rened devices such as silicon detectors and gas chambers can be easily used. The detector

of the Columbia University microbeam consists of a pulsed ion counter lled with P10 gas and

built into the microscope objective used for the cell localization (Randers-Pehrson etal., 2001),

allowing simultaneous viewing of the targeted samples and monitoring of the particles delivered.

Such congurations, however, require that the delivered particles have enough energy to pass

through the samples and escape with sufcient energy to be accurately detected, setting a limit on

the low energy usable. In many cases, it is also necessary to remove the culture medium, requir-

ing additional procedures (such as humidity control devices) to keep the cells viable during the

irradiation process.

Finally, an interesting approach has been adopted at the GSI, where the use of heavy ions

makes placing the detector either above or below the sample less than optimal. In order to pre-

serve the quality of the focused beam without imposing strong constraints on the sample holder,

Heiss and colleagues (Heiss etal. 2006) used an in-vacuum channeltron to collect secondary

electrons produced by the heavy ions crossing the Si

vacuum window that was coated with

a thin layer of 50 μg/cm

CsI to increase the secondary electron emission. As all the detection

elements are in vacuum, this conguration has the advantage of minimal interference with both

the

radiation beam and the sample holder/imaging apparatus. A detection efciency of 99.5% for

carbon ions has been claimed.

21.5.3 beaM orientation and SaMple holder

One of the major difculties in developing a microbeam facility for radiobiological studies is to

create and maintain suitable conditions for the biological sample for the duration of the irradiation

experiment. Despite sounding trivial, this is very complex considering all the other requirements

and constraints imposed by the radiation delivery, detection, and cell imaging. Considerable efforts

have been devoted to developing appropriate sample holders that, while supporting the cells without

causing extra damage/or stress, would not interfere signicantly with the radiation beam. The main

requirements for a microbeam sample holder are the following: (1) the ability of samples to rmly

Application of Microbeams to the Study of the Biological Effects of Low Dose Irradiation 583

attach to the holder to achieve accurate individual targeting of multiple targets can only be achieved

for rmly attached samples, (2) the property of being optically transparent and non-UV uorescent

(see Section 21.5.4), (3) the property of minimum interference with the radiation beam and easily

accessible on both sides by the radiation probe and the optical devices. This has led to the popular

choice of thin (<5μm) polyester lms (Mylar or prolypropylene) glued or clamped into conventional

or custom-made Petri dishes on which the cells are seeded and through which the radiation is deliv-

ered. Other materials used as a cell substrate are 100s nm thin Si

membranes and track plastic

detectors (CR39 and LR115) etched to submillimeter thickness. The latter also offer the option of

visualizing the track traversals and relating them to the sample positions. Although this option can

be used to check the number and site of traversals, it cannot be used to predetermine the number of

particles delivered. The sample holder design is however strongly inuenced by the orientation of

the radiation probe. While microbeams expressly designed for radiobiological studies tend to have a

vertical beam, most of the existing facilities are horizontally oriented, being adapted from existing

accelerator beamlines. Taking advantage of gravity, vertical microbeams can generally adopt sim-

pler sample holders to keep cells in an optimal environment for a long period of time. By contrast,

horizontal beams require more complex designs to either enclose cells in a leak proof chamber that

may limit or preclude access to the sample for the objective/detector, or control the humidity level

with

medium ow or humidied gas. This can limit these to short irradiation times.

21.5.4 opticS and SaMple preparation

In order to irradiate individual biological targets with high accuracy, the samples need to be iden-

tied with high resolution. Microbeam facilities, therefore, make effective use of state-of-the-art

imaging devices and techniques. The most common approach is to use commercially available

microscopes (often custom modied to be integrated into the radiation setup) and uorescence

microscopy. By taking advantage of the variety of molecular uorescent probes available, subcel-

lular targets of interest are labeled using specic uorescent dyes and subsequently viewed in situ

following snapshot UV illumination. Images are usually collected using CCD devices and analyzed

by appropriate software for target identication and coordinates recording. Using this method, very

fast and accurate target localization can be achieved due to the excellent quality of the images col-

lected and the ability to perform such an operation immediately before or during the irradiation

procedure. Despite using very low concentrations of dye (down to the μM range (Randers-Pehrson

etal., 2001)) for minimum incubation periods (<1h) and short UV exposures (<100ms (Folkard

etal., 1997a,b)), the inuence of these factors on the cellular response cannot be completely elimi-

nated, and accurate control (sham irradiation) experiments are necessary to check their negligible

contribution and correct for it. Typical dyes used for microbeam experiments are the DNA binding

compound Hoechst 33258 (is excited at 350nm and emits at 461nm) for the nucleus and Nile Red

(excites at 485nm and emits at 525 nm) for the cytoplasm. However, the advent of green uorescent

proteins (GFP) approaches has redened the need for uorescent microscopy. GFP can be expressed

in different biological structures enabling morphological distinctions to be made, and as they are

usually much less harmful when illuminated in living cells, they offer great potential for automated

live cell uorescence microscopy and microbeam studies.

An alternative approach using phase-contrast microscopy technique has also been exploited suc-

cessfully by the INFN-LNL laboratories (Gerardi etal., 2005). The great benet of this method is

the lack of external insults for biological samples as only visible light is used. However, in order to

obtain high quality images easily processable by software, phase-contrast microscopy requires the

illumination source and microscope objective to be placed at opposite sides of the samples. This is

usually a challenging issue for microbeam facilities as physical constraints on one side of the sample

are imposed by the radiation delivery setup. To overcome such limitations, the INFN-LNL adopted

an off-line sample recognition procedure that relies on a sophisticated translation of the sample

holder

to match the target coordinates with the radiation probe.

584 Charged Particle and Photon Interactions with Matter

The ability to quickly and accurately locate biological targets is also critical for performing

good microbeam experiments. The speed of image acquisition and processing is usually the deter-

mining factor inuencing the cell throughput. Using relatively low microscope objectives (×20

to ×40) and relying on good CCD devices, large eld images of samples are generally acquired

with a resolution of the order of 1 μm. This allows irradiation throughput in the range 5,000–

15,000 cells/h when supported by automated cell recognition algorithms. Cell density and dose

rate also inuence the nal throughput. Experiments with yields as low as 10

−4

(i.e., mutagenic

and oncogenic end points) are therefore possible. It has to be noted that the accuracy with which

biological targets are identied is just as important as other aspects of the microbeam experiment.

Depending on the end point investigated and the experimental design, considerable uncertainties

can be introduced if not all viable targets are correctly identied. These include missing viable

cells, generally caused by poor image quality due to nonuniform UV illumination or dye uptake

and/or inability to resolve two or more samples in close proximity, and mistaking background

noise for biological targets.

21.5.5 poSitioning Stage

The accuracy to which samples can be aligned with the radiation probe has to at least match the

spatial resolution of the beam. This is generally achieved by using automated translation stages

with submicrometer precision and high reproducibility. Stage speed also plays an important role

in the scanning and location of the target samples, contributing to the nal irradiation throughput

achievable.

21.6 x-ray miCrobeams

Following the success of charged-particle microbeams, x-ray and electron microbeams have also

started to be developed providing quantitative and mechanistic radiobiological information that

complement charged-particle studies. Unlike charged particles, soft and ultrasoft x-rays (i.e.,

∼100s eV–10s keV photons) interact almost exclusively by photoelectric effects and therefore

don’t suffer from scattering issues. Higher energy x-rays (>10 keV) are not suitable for microbeam

applications as the range of secondary electrons produced by the photon absorption is greater

than the x-ray spot resolution, invalidating the efforts made to reduce the radiation beam to

micron or submicron size. Even if 10keV x-rays are focused into a micron spot inside a cell, the

range of the secondary electrons produced will spread the dose delivered over a volume consider-

ably larger (the range of 10 keV electron in tissue is ∼2.5 μm). While extreme care has to be taken

for charged-particles microbeams to ensure that the radiation beam accuracy is not compromised

by scattering through the collimator, vacuum window, sample holder, and/or detector devices, it is

in theory possible, with x-rays, to achieve radiation spots in air of an order of magnitude or more

smaller than those so far achieved with ion beams. Moreover, such high spatial resolution is main-

tained as the x-ray beam penetrates through cells, making it possible to irradiate targets that are

several tenths or hundreds of microns deep inside the samples with micron and submicron preci-

sion. The lack of scattering and the superior spatial resolution offered by x-ray microbeams make

them the perfect tool to investigate questions regarding the role of subcellular targets in 2D cell

cultures as well as in more complex 3D cell structures. Finally, another critical role covered by

x-ray (and electron) microbeams concern the biological effectiveness of different LET radiations.

It is commonly accepted that radiation effects such as apoptosis, mutation, and chromosomal

instability are highly LET dependent (Lorimore etal., 1998; Kadhim etal., 1992). Electrons pro-

duced by the absorption of soft x-rays will produce ionization patterns that closely resemble the track

structure of a low-LET irradiation in a way that is not possible to achieve using charged particles.

In this respect, x-ray microbeams complement the work done with charged-particle facilities to

investigate the LET dependence.

Application of Microbeams to the Study of the Biological Effects of Low Dose Irradiation 585

21.6.1 x-ray Source

Generally, x-ray microbeams are based on similar concepts elucidated earlier for the charged-

particle facilities; however, they face different technical issues due to the radiation production

and the different nature of its interaction with matter. Some of the approaches described above for

sample preparation, holding, imaging, and positioning are also applicable for x-ray microbeams.

However while charged-particle microbeams require the use of particle accelerators, x-ray micro-

beams can be developed using table-top sources (Folkard etal., 2001). Clear advantages of such an

approach are the reduced costs, the simpler operation (experiments can be performed by a single

scientist), and the possibility of building a compact facility. Compact table-top x-ray sources that

have been used for microbeam facilities have been of the electron bombardment type, where the

electron beam generated by a conventional electron gun is accelerated up to 20–30 kV and focused

onto a solid target producing a “point-like” x-ray course (see Figure 21.2). The electron beam is

generally produced by heated tungsten laments (or more recently using single crystal laments

with brightness above 1 × 10

A/cm

sr) and focused either by permanent magnets or by electro-

static lenses. Despite electron bombardment x-ray sources being an established technology, the

requirement of focusing a high electron current into a small spot (in order to achieve a bright x-ray

point source) is not easily achievable and still represents a major limitation for x-ray microbeams.

Currently the best performances have been achieved by the Queen’s University x-ray microbeam

where up to 1 mA of 15 keV electrons have been squeezed into <20 μm diameter spot. The radiation

spectrum generated by electron bombardment sources is composed of Bremsstrahlung as well as

characteristic x-ray lines of the target element upon which the electrons impinge. In order to obtain

a monochromatic x-ray beam, which is required both for optimum working of the x-ray focusing

devices as discussed below and for a correct interpretation of the biological effects caused, the

Bremsstrahlung component is removed either by ltration and/or by reection. Depending on the

x-ray energy of interest, low energy Bremsstrahlung can be removed by using appropriate lters

while the high-energy component is usually eliminated by using a reection approach. Taking

advantage of the total reection phenomenon that occurs for low energy photons below a critical

angle, highly polished at surface mirrors can be used to attenuate the high-energy component

X-ray focusing

optics

Bremsstrahlung

filtering

mirror

Focusing magnet

Electron gun

Solid target

Figure 21.2 The x-ray source of the Queen’s University x-ray microbeam, showing the key elements of an

electron

gun, focusing magnet, the target chamber, and the x-ray focusing optics.

586 Charged Particle and Photon Interactions with Matter

while reecting a signicant fraction of low energy x-rays. The nal result is the production of a

nearly monochromatic x-ray beam corresponding to the brightest characteristic x-ray line of the

electron bombarded target. As mentioned, the major limitation of the electron bombardment x-ray

sources is represented by their limited brightness and size. The spot size into which it is possible to

focus electrons is ultimately limited by the electron beam current as space charge effects will pro-

duce a repulsive force that will compromise the focusing action of permanent or electromagnetic

lenses. Moreover, the x-ray production by electron bombardment is not a very efcient process with

only a small percentage of the electron beam energy actually transformed into x-ray production.

Most of the energy deposited by the electrons into solid targets will be dissipated as heat. Special

care, therefore, is to be taken when designing solid targets to avoid high temperature damage.

Using xed thick targets, it is generally necessary to compromise between high electron currents

and the size of the electron focus as the maximum power that can be possibly dissipated on a thick

target depends on the power density of the impinging beam (i.e., electron current × accelerating

electron voltage/area electron focus). Cooled or rotating targets are other approaches that have been

investigated to increase the electron current and therefore the x-ray production rate. However, both

approaches present technical issues for their implementation in microbeam facilities. Cooling the

target is an efcient method but only if the heat extraction device (generally water or gas) can be

placed very close to the electron spot. As the electrons are focused into a micron sized spot and

have limited penetration in the target, it is basically impossible to efciently cool the area of the

target where the electrons impinge. Such an approach will only bring minor improvements in terms

of x-ray production. Rotating targets offer more promising benets and have indeed been used to

produce x-ray sources for microbeam applications (Lekki etal., 2009). Higher electron currents

with higher voltages and small spot sizes (∼1 μm) have been reported. However, due to the rota-

tion of the target, this type of x-ray source generally suffers from mechanical vibrations that may

ultimately affect the resolution and accuracy of the nal x-ray focused beam. The smallest x-ray

spot size achieved using a rotating target x-ray source is 7 μm. Another limitation of the electron

bombardment sources is the range of x-ray energies that can be possibly generated. X-ray energy is

determined by the energy of the electrons and the target material. In order to produce high-energy

x-rays, high Z materials and high-energy electrons have to be used. While high-energy electrons

are harder to focus and dissipate higher power on the target, high Z materials offer lower cross sec-

tion values for the x-ray production. Both factors conspire against the production of suitable mono-

chromatic high-energy x-ray beams and, to date, x-ray microbeam facilities have successfully been

developed only for energies up to 4.5keV (characteristic K

titanium line). Some of the electron

bombardment source limitations may be overcome using proton bombardment sources. The advan-

tage of using protons instead of electrons lies in the higher x-ray production cross section (i.e.,

more energy is transformed into x-rays) and the reduced (in some cases negligible) Bremsstrahlung

component that does not require beam ltration. The use of protons, however, requires linear accel-

erators and although higher proton beam currents can be achieved and focused into smaller spot

sizes, the maximum power that can be possibly dissipated into the target will ultimately limit their

applicability for microbeam systems.

Synchrotron radiation generated by high-energy electrons accelerated in storage-type circular

accelerators also provides a very interesting source of x-rays for microbeams applications. Despite

the difculties in working with a horizontal beam line (see Section 21.5.3), the high brightness, wide

energy range, and nearly parallel directional beam of synchrotron facilities offer great potential

from both the technical development aspect as well as for radiobiological applications. The nearly

parallel and monochromatic beam assures that the best focusing resolution can be achieved by

the employed x-ray optical devices, while the tunable x-ray energy is an incredibly powerful tool

to address critical radiobiological phenomena. However, the cost of running a synchrotron-based

microbeam and the restricted beamlines available are considerable limiting factors, and so far the

only microbeam facility taking advantage of synchrotron light is based at the Photon Factory in

Japan (Kobayashi etal., 2009), used to irradiate biological samples with a 5 μm x-ray beam of

Application of Microbeams to the Study of the Biological Effects of Low Dose Irradiation 587

4–20keV. Finally, it has to be mentioned that laser plasma or electrostatic pinch plasma sources

(Blackborow etal., 2007) are also being considered as possible x-ray sources for microbeams. Such

sources are characterized by small size and high brightness and may provide valuable input in

microbeam

development.

21.6.2 x-ray opticS

The rapid increase in the development of x-ray microbeams is due to the high spatial resolution

possible, with which to focus x-rays. Sophisticated x-ray optics have been under development world-

wide for many years for lithography and astrophysics applications. Microbeam facilities can there-

fore

rely on a considerable pool of expertise and technical devices.

X-ray

beams can be focused using one of two different principles: reection or diffraction.

Focusing methods based on reection use curved surfaces coated with a single layer (mirror) or

several layers (multilayers) to bend the x-ray beam and force it to converge in a single spot. They can

be used for a wide energy range, have large acceptance (i.e., can collect a large fraction of the x-ray

beam), and benet from high efciencies even for high-energy x-rays (>10keV). The most com-

monly used conguration is the so-called Kirkpatrick–Baez design where two orthogonal parabolic

mirrors are used to successively reect the x-ray beam, achieving a two-dimensional focused spot.

This approach provides high accuracy and exibility as it offers the possibility to vary the focal

distance and focus different x-ray energies by “simply” changing the mirror’s orientation to alter the

x-ray incident angle. However, the mirror alignment/tuning is a very critical step usually performed

by automated precision motors to optimize the eight degrees of freedom of the system. The coalign-

ment of many reecting surfaces to form an optimum image is a difcult process as aberrations are

easily introduced and the nal focusing accuracy compromised. Another factor limiting their use

for microbeam applications is the typically long focal distance required by the reecting mirrors to

produce

optimum results.

The

majority of the x-ray microbeam facilities, however, employ diffraction techniques to bring

the x-ray beam to a ne submicron spot. Fine x-ray probes (<50 nm diameter) are achieved in x-ray

microscopy using Fresnel zone plate lenses that are circular diffraction gratings with increasing

line density (see Figure 21.3). Being a diffraction device, zone plates will focus the incident x-ray

radiation into a series of orders progressively closer to the zone plate itself. In order to stop all but

the rst order focusing from reaching the sample, an arrangement of masks have to be accurately

aligned. The mask is usually an axial disk mounted on the zone plate itself and an equal size pin-

hole. The rst order diffraction focus is generally chosen as it is the most intense, while the zero

(i.e., undiffracted) and secondary orders have to be removed as they would reach the sample unfo-

cused, therefore spoiling the beam resolution. Due to diffraction principle constraints (i.e., gap

between transmitting and opaque zones needs to be of the order of the wavelength of the incident

x-ray radiation), zone plates are very small devices (a few hundreds of microns in diameter) with

ne structures down to 50 nm. Combined with their low focusing efciency (theoretical max for

low energy x-ray ∼20% and rapidly decreasing at higher energies), this means that zone plates are

only able to focus a very small fraction of the available x-ray beam. Despite these limitations, the

superior spatial resolution offered by the zone plates and the relatively straightforward applica-

tion make them particularly suitable for microbeam applications (Folkard etal., 2001; Thompson

etal., 2004).

Both focusing methods described above require a parallel incident x-ray beam in order to pro-

duce the nest possible spot. In case of non parallel beams, zone plate and mirror work can be

compared to that of a normal optical lens; they basically demagnify the x-ray source. In practice,

the nal x-ray spot is therefore determined by the size of the x-ray source and its distance from the

focusing device rather than the focusing device itself. Another requirement unrelated to the focusing

approach is that microbeam x-ray sources be point-like and bright enough to illuminate the optical

device

with suitable photon uxes.

588 Charged Particle and Photon Interactions with Matter

21.6.3 x-ray detection and doSiMetry

Due to the nature of the x-ray interaction with matter, it is not possible to detect the same photons that

have or will interact with the biological samples. For x-ray microbeams, the irradiation is basically

a time exposure determined on pre-irradiation characterization of the emerging beam. The focused

x-ray beam is usually characterized using an x-ray detector (proportional chamber or photodiode)

operating in a photon counting mode and scanning a knife edge mask across the beam to determine

the prole and its location. Considering the low dose rate generally available for x-ray microbeams

(fraction of Gy/s) and the large number of photons necessary to deliver a suitable dose (for carbon K

x-rays ∼10,000 photons are necessary to deliver 1Gy to a typical mammalian cell), a fast mechanical

shutter is appropriate to assure an accurate dose delivery and uncertainties due to the variation in the

ux of photons generated (which follows a Poisson distribution) is negligible. The methods used to

report the dose absorbed by the samples exposed to localized and/or partially penetrating radiations

are often the basis for discussion and it may present problematic issues such as the critical mass,

which should be considered for dosimetry calculations. In the case of charged-particle microbeams,

the number of particle traversals is usually reported as well as the average nuclear dose, calculated

considering the energy deposited and the entire cell nucleus mass. For x-ray microbeams, the num-

ber of photons that deposit energy inside the samples cannot be precisely estimated, and despite the

high non homogeneous energy deposition (especially for ultrasoft x-rays that are highly attenuated),

the average nuclear dose is usually reported. It must also be noted that cell morphology is a critical

parameter, and precise information on the cytoplasm and cell nucleus thickness must be taken into

consideration when assessing the effects induced by x-ray microbeam irradiation.

21.7 eleCtron miCrobeams

Charged-particle microbeams cover predominantly the high LET spectrum of ionizing radiation.

The energy of x-ray facilities is still too high to be classied as low LET, and therefore, the ideal

way to produce a low-LET microbeam is to use a fine beam of energetic electrons (<100 keV).

15 kv

10,000×

1.00 μm

OSA

0063

Figure 21.3 Schematic illustration of the focusing principles of a zone plate—An order sorting aper-

ture (OSA) prevents alignment with all but the rst order diffraction prevented from reaching the sample.

A detailed example of a zone plate lens is also shown.

Application of Microbeams to the Study of the Biological Effects of Low Dose Irradiation 589

Asfor x-ray, electron microbeams are generally self-contained units with relatively low development

and maintenance costs. They rely on standard electron guns and electrostatic devices to produce

and accelerate energetic electron beams that are subsequently reduced to micrometer size by the use

of collimators (generally laser-drilled apertures) or electromagnetic focusing. The use of collimat-

ing apertures is often employed (Sowa etal., 2005) due to their straightforward application. As the

electron penetration is very limited, only thin apertures are required, making the alignment process

much easier than what was discussed earlier for charged particles, and the low beam uency is suit-

able for low-dose experiments. However, since electrons are easily scattered as they exit the vacuum

system, the distance and the amount of material between the collimator and the samples have to be

minimized. As electrons also scatter as they interact with the biological samples, it is impossible for

electron microbeams to achieve targeting resolutions of the micron or submicron level despite the

actual size of the focused beam. Microdosimetry studies (Wilson etal., 2001) indicate that the

energy deposited by a 25keV electron track is barely contained within a typical mammalian cell and

the ionizations are nonuniformly distributed across its nucleus. As the energy increases, the electron

range also increases, but the later component of the scattering actually reduces in favor of the for-

ward component. The maximum later scattering is calculated to occur at ∼50keV when about 16%

the energy deposited in the targeted cell leaks into the neighboring samples (Miller etal., 2001).

A great advantage of the electron microbeam is also the ability to easily vary the electron energy

(and therefore the LET) in order to investigate the relative biological importance of various parts of

the electron track. Using electron energies ranging from 25 to 80 keV, Sowa Resat and colleagues

(Sowa Resat and Morgan, 2004) showed that the peak in micronuclei formation is shifted to lower

doses as the electron energy decreases. This has been successfully used to strengthen the hypothesis

that the electron track end is mainly responsible for the induction of biological damage. Finally, for

more energetic electrons, considerable penetration can be achieved (path length of 90keV electron

exceeds

100

μm)

allowing irradiation of thick tissue samples of 5–6 cell layers.

21.8 Future developments

The interest in the microbeam technique is such that the number of facilities developed worldwide

has increased from the initial three of the mid 1990s to over thirty acknowledged in 2006. The great

contribution to radiobiology offered by precise and accurate irradiation of the samples justies the

efforts devoted by several laboratories in developing new methodology and indeed improving the

existing microbeam facilities. Some of the future developments that are already under investiga-

tions are reported below.

21.8.1 beaM Spot Scanning

One of the great advantages offered by the electrostatic focusing of charged particles over the col-

limator approach is the possibility to precisely deect the radiation beam. Accurate beam spot scan-

ning can be achieved by simply varying the voltage of the electrostatic lenses and moving therefore

the radiation probe to the sample position in contrast to what is currently done. As electrostatic

beam movement can be done much faster than mechanical stage movements, this approach has

the potential to speed-up the irradiation process and signicantly increases the cell throughput.

However, radiation beams can only be scanned over very small distances without losing focusing

accuracy, limiting the exploitable area of the sample holder. A practical solution would be to couple

the electrostatic beam scanning to a coarse stage movement in order to be able to use a larger sample

area. Finally, the beam scanning irradiation approach necessitates accurate calibration procedures

to determine the correct parameters of the electrostatic lens necessary to move the beam to the

desired sample position. While beam scanning is an established technique, the submicron precision

required by a microbeam experiment represents a technical issue that has prevented such techniques

be successfully employed so far.