A free-electron laser~FEL! is a free beam of relativisticelectrons that passes through a periodic magnetic fieldresults in the stimulated emission of light.1 As such, theseaccelerator-based devices convert the kinetic energy ofelectrons into light. In contrast, the medium for conventiolasers is bound electrons and the accompanying atomicclei, which restrict laser performance in some ways. FEcan be quasicontinuous or have high-peak and in some chigh-average optical power. They can produce long optpulses, produce pulses with durations as short as subpicoonds, or have a complex superpulse structure. The detering factors that establish the detailed characteristics oflight emitted by a FEL include the electron beam energy aelectron pulse structure as well as the magnetic field chateristics. FELs are continuously tunable and have succfully operated in the microwave, far infrared, midinfrarevisible, ultraviolet and x-ray ranges. There are several exlent reviews of FELs.2
There are many FELs worldwide located in relativelarge research facilities, each typically providing thousanof hours of beam time. Leading centers for biomedical abiophysical FEL applications include the far-infrared faciliat the University of California, Santa Barbara~UCSB!, mid-infrared facilities at Duke, Stanford, and Vanderbilt Univesities, The Jefferson Laboratory, FELIX in The Netherlanand laboratories in France and Japan, and ultraviolet faties at Duke and the LURE laboratory in France. Progrcontinues in FEL physics, in particular, the pursuit of utrashort pulsed x-ray FELs.
Multidisciplinary research teams have pioneeredearly applications experiments, taking advantage of
a!Author to whom correspondence should be addressed; [email protected]
at
helu-ssesalec-in-edc-s-
l-
sd
,li-s
ee
unique light source capabilities. As they have proved scessful, experimental techniques have migrated betweenphysical, biological, and biomedical sciences. Here we wreview research accomplishments in the biophysical andmedical sciences as well as comment on future prospeNaturally there have been numerous investigations of Fapplications in the physical sciences,3 but they fall outside ofthe scope of this review.
FELs are versatile light sources that allow pioneeriapplications research by tuning to wavelengths of chowith relative ease. Many FEL laboratories are centers ofcellence for interdisciplinary research where scientific pnomena have been discovered and characterized. Thisunderstanding has allowed the specification of novel, decated technology to exploit these phenomena to improvehuman condition. The spin-off technology is less compand resource demanding than FELs, but indeed is a coquence of FEL development and FEL-applications resea
II. TISSUE ABLATION WITH THE MARK-III FEL
The argon ion, excimer, and CO2 lasers are in routinemedical use.4 The visible argon ion laser has the advantathat it can be transmitted through the structures of theand focused to coagulate~remodel! and thus treat tissueHowever, visible lasers are not effective tools for tissuelation, i.e., removing a targeted volume of tissue such tthe surrounding tissue is biologically viable. In contrast, tultraviolet ~UV! excimer has the advantage that it can etissue with essentially no collateral damage. Consequethe excimer is a very effective tool for reshaping the cornto correct its optical properties, i.e., laser vision correctioTypically, the outer surface of the cornea, a cellular tissuepartially cut and lifted to expose the interior stroma, an aclular tissue, which is laser etched. Concerns about the potial mutagenic effects of ultraviolet radiation, i.e., photchemical effects mediated by excited electronic states, hil:
largely prevented medical applications of the UV to cellutissues. The infrared spectrum is typified by vibrational trasitions, with far less concern about photochemistry. Thefrared CO2 laser is used for surface treatments of some tistypes, where the ablated volume is surrounded by a zondenatured and possibly charred tissue, useful for controlbleeding. In these applications the collateral damage isceptable since the overall tissue remains biologically viabHowever, there are many potential medical applicatiowhere such a zone of collateral damage is unacceptaConsequently, there has been great interest in the biomecommunity to explore the ablative properties of infrarlasers5 to see if they can mimic the clean cutting of thexcimer.
The Mark-III FEL is a tunable, infrared source in th2–10 mm range with high-peak and high-average powe6
The Mark-III produces a superpulse: the ‘‘micropulse’’about a picosecond in duration and contains tens of mijoules; the ‘‘macropulse’’ is a train of tens of thousandspicosecond pulses with a duration of about 5 ms and delivtens of millijoules; the repetition rate of the macropulse isto 30 Hz.
A. Infrared wavelengths and thermodynamics
Tissue typically is about 75% water, which is strongabsorbed in the infrared. So strongly absorbed that transsion infrared spectroscopy is not a viable option for invegating tissue in its native state of hydration. However, atteated total reflectance~ATR! is a near-field or evanescenwave technique ideal for strongly absorbing samples. TATR sampling technique is particularly powerful whecoupled with Fourier transform infrared~FTIR! spectros-copy. In practice, the hydrated tissue is placed on the surof an infrared material with a relatively high refractive indeBroadband infrared light propagates in the material suchit reflects from the material/tissue interface under the contion of total internal reflection, i.e., the tissue is only exposto the evanescent wave. Consequently, the highly absorsample is optically sampled in the near field and we gainof the advantages of FTIR spectroscopy.7
Figure 1 presents a midinfrared spectrum of cornea, nral tissue, and dermis. To first order, the midinfrared msures the localized vibrational modes of tissue componeThus these spectra are similar because the midinfrared isparticularly sensitive to higher structural organization. Cosequently, the results summarized below can be generato many tissues. These spectra serve as guides for seleFEL wavelengths for investigating tissue ablation as wellproviding a biophysical foundation for interpreting the rsults.
A series of experimental investigations have demstrated that targeting a midinfrared Mark-III FEL to wavlengths near 6.45mm results in tissue ablation at a substatial ablation rate with minimal and at times undetectacollateral damage. The seminal study reported investigatof FEL ablation in ocular, neural, and dermal tissues aproposed a thermodynamic model to account for theseperimental observations.8 Wavelengths near 6.45mm coupleinto the spectral wing of the bending mode of water cente
r--eofgc-.
sle.cal
o-frsp
is---
e
ce
ati-dngll
u--
ts.ot-edings
-
-
nsdx-
d
at 6.1mm as well as the amide-II vibrational mode centerat 6.45mm, both relatively broad spectral features. Thermdynamic reasoning suggested that the reduction in collatdamage is due to differential absorption; more specificatissue integrity is compromised due to laser heating ofnonaqueous components of tissue prior to explosive vaization due to laser heating of the aqueous compone
FIG. 1. ATR-FTIR spectra of cornea, neural tissue, and dermis in nenative states of hydration. The dominant feature is the OH-stretch modwater near 3300 cm21 ~3 mm!. The partially resolved spectral band ne1650 cm21 is deconvolved in the inset. Three modes are found to contribthe amide I vibrational mode of protein at 1665 cm21 ~6.0 mm!, the OHbending mode of water at 1640 cm21, and amide II vibrational mode ofprotein at 1550~6.45mm!.
a
ith
at
hanoI
nt
ito
d
vetaer
aratdpe
smd
live
entabla-ily
dia-Hz
ionsticllel
o-e totionsticthemei-par-
L-
ofghsedol-wase-
forin-i-
de-ver,dov-
eti-
om-la-i-of
edersveral
su-hisayun-dent
ra-di-tle,setat-
, an
3210 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
These observations laid the groundwork to pursue humsurgery with a Mark-III FEL.
Experiments that investigate the ablation of bone wthe Mark-III FEL demonstrate that the 6.1–6.45mm range isefficient in cutting cortical bone.9 There is relatively littlecollateral damage and the healing response proceedsfaster rate during the first two to four weeks comparedcutting with a bone saw. The investigators concluded tinfrared ablation in bone is due to explosive vaporization athat ablation is enhanced by targeting the protein matrixcortical bone via the overlapping amide I and amidemodes in the unresolved spectral band in the 6.1–6.45mmrange. Experiments that investigate the ablation of dehard substances between 5 and 12mm, including phosphatebands at 9.5mm associated with hydroxyapatite, exhibwavelength dependent surface modifications ranging frpartial vitrification to surface roughening.10 Additional inves-tigations of Mark-III FEL ablation of other soft tissues anmaterials have been reported.11
B. Pulse structure and dynamics
While the thermodynamic model accounts for the walength dependence of infrared tissue ablation, it lacks deof the dynamics. The time constants of the Mark-III suppulse~Fig. 2! are 1 ps, 350 ps, 2–6ms, and tens of ms andcorrespond to the micropulse duration, micropulse seption, macropulse duration, and macropulse repetition rrespectively. In principle, these time constants corresponmultiple dynamic processes that need to be sorted out exmentally.
A full macropulse that delivers tens of millijoules drivemultiple dynamic processes. In an effort to reduce the nuber of processes, a broadband infrared Pockel’s cell wasveloped based on a CdTe crystal.12 The Pockel’s cellswitches out a train of micropulses as short as 60 ns, deering hundreds of microjoules, and as long as 2 ms, deli
FIG. 2. Superpulse structure from a Mark-III FEL. From top to bottomtrain of three macropulses, a single macropulse, three micropulses, asingle micropulse.
n
t aot
df
I
al
m
-ils-
a-e,tori-
-e-
v-r-
ing 1 mJ, enabling pulse duration and wavelength dependmeasurements of the FEL-induced stress transients andtion plumes.13 The back surface of a gelatin sample readadheres to the surface of a 9mm thick PVDF piezoelectricfilm that measures stress transients induced by FEL irration. The stress transients were measured with 500 Mbandwidth electronics. For fluences below the ablatthreshold, the superposition of asymmetric thermoelawaves is detected. A continuous HeNe beam that is parato and 10mm above the front surface of the gelatin is fcused on a silicon photodiode with a 25 ns response timmonitor the ablation plume. For fluences above the ablathreshold, the superposition of an asymmetric thermoelawave with momentum recoil is detected. Furthermore,wavelength dependence of the duration of the ablation pluindicates that 6.45mm radiation compromises the mechancal properties of the gelatin, as do measurements of theticle size distribution in the ablated material.14
Dynamic light scattering has been used to monitor FEinduced denaturation of cartilage and cornea.15 The tissuesamples are about 1 mm thick and initially in a statenearly native hydration. Light of 600–700 nm passes throua sample from the back surface, such that the light is focuon the front surface. Nearly forward scattered light is clected with a photodiode array, where the capture rate1–4 Hz, while the front surface is radiated by the FEL. Ninteen wavelengths between 2.2 and 8.5mm were investigatedto determine the threshold energy and kinetic coefficientsdenaturation. The signature for structural alteration is ancrease in the distribution of visible light scattering. The minmum threshold was observed at 6.45mm. For most FELwavelengths, there is an inverse correlation between thenaturation threshold and the absorption coefficient. Howefor wavelengths near 3 and 6mm, the denaturation thresholdoes not obey this inverse correlation and instead is gerned by heating kinetics.
Experimental attempts to ablate tissue with a kHz reption rate, picosecond optical parametric amplifier~OPA!,where peak intensities and total energy delivered were cparable to those typically used in Mark-III FEL tissue abtion, were unsuccessful.16 Consequently, a detailed theoretcal investigation of the role of nanosecond dynamicsthermal diffusion in infrared tissue ablation was carriout.17 It was found that the temperatures of the surface layof tissue water reached many hundreds of degrees in senanoseconds: at this rate the outer saline layers becomeperheated and lead to explosive vaporization. During ttime, the temperature of the surface layers of protein meither exceed or trail the water temperatures by tens to hdreds of degrees, depending upon the wavelength-dependifferential absorption of water and protein. At 6.45mm theprotein temperatures uniformly exceed the water tempetures and a Arrhenius treatment of protein dynamics incates that ductile, native protein begins to convert into britdenatured protein. Apparently the brittle fracture at the onof explosive vaporization leads to the confinement of colleral damage.
A key to the dynamics governing infrared tissue ablatis the differential heating rates for protein and saline. Tdynamics are not established by the superpulse structurese. Instead, the heating rates must be rapid enough to dthe separation in temperature between protein and waterers to achieve both protein denaturation and the superheof water. This observation leads to reconsideration of theof the Mark-III FEL superpulse structure in tissue ablatioIn particular, theoretical calculations indicate that pulseration of tens of nanoseconds with the same average enas the superpulse should achieve similar ablative results
Little is known about the polymer dynamics that leadbrittle fracture. In particular, what is the wavelength depedence for achieving mechanical confinement in additionthermal confinement on such short length and time scaleseems evident that linear and nonlinear pressure wavesplay a role in the onset of nonthermal collateral damageaddition, it has been proposed that the Mark-III FEL, ffluences below the ablative threshold, can operate as a swave generator.18 Since the pressure waves alter membrapermeability, there may be drug delivery applications.
Within the macropulse, the Mark-III FEL is a GHz repetition rate picosecond laser with high-peak and higaverage power.19 This is the consequence of the pulse powsystem, including the long-pulseS-band klystron, the modulator, and the pulse-forming network.20 FELs are lightsources with unique capabilities that are typically compand expensive. Relaxing the constraint from a GHz repetirate, tunable infrared laser to a single wavelength, lower retition rate, infrared laser with a pulse duration of tensnanoseconds is technologically much more forgiving. Tanalysis suggests the feasibility of a nonaccelerator-bamedical laser, with a pulse duration of 10 ns, operating n6.45mm.
III. MARK-III FEL AS A SURGICAL LASER
The Vanderbilt Mark-III FEL first lased in 1991 and intially served as a research tool for five years. The advenhuman medical applications necessitated upgradingMark-III to meet the operational standards of a medical lawhich are quite different from the standards of a resealaser. In 1996, a major effort commenced to upgradeMark-III FEL to the status of a medical laser for humasurgery.20
A. Failure analysis
The types of failures for the FEL and its subsystemsthe gamut from those that self-correct, like an arc inklystron tube, to those that are nearly catastrophic and mrequire days to weeks to repair, e.g., vacuum leaks orfailure of high voltage components in the pulsed power ssystem. The failures fall into the following categories:
~1! major, planned shutdowns scheduled typically more thone month in advance and extending over more thafew days;
~2! routine maintenance, scheduled typically at least a win advance and extending over a few hours;
eer
ivey-
ingle.-
rgy
-oItayn
esse
-r
xnp-fsedar
ofer,his
neaye-
na
k
~3! nonroutine maintenance that requires the FEL to be sdown for more than 1 h or so, butthat can be scheduleat the beginning of the next running day;
~4! repair of minor, immediate failures that may require miutes to an hour;
~5! repair of medium size failures that may require a fehours to a day;
~6! repair of major failures that requires many daysweeks;
~7! self-correcting faults such as klystron arcs which intrupt the operation for seconds or minutes.
Obviously the impact on surgeons using the FEL dpends on the category of failure. Interruptions of a few mutes have negligible impact for a surgeon performing a 3 hprocedure, whereas catastrophic failure would disrupt patcare. Predictability is essential for human surgery. PredicFEL performance one day into the future based on currbehavior and accumulated diagnostics was one of the mgoals of the reliability upgrades.
The reliability plan included~1! stocking spare partsespecially klystrons,~2! upgrading the control system foboth better robustness and for more and better diagnosand ~3! following sound engineering practices in the higvoltage, pulsed-power systems.
B. Subsystem upgrades
The klystron is a long pulse, 30 MWS-band microwaveamplifier manufactured by Triton ETD. One such klystrohad provided eight years of operation, but was beginningshow poor high-power performance. Klystrons have retively high infant mortality and require several weekscommission. Consequently at least one fully conditionspare is now kept on hand. The modulator includes a 406 A dc power supply that charges up a bank of capacit@pulse-forming network~PFN!# and switches the PFN acrosa transformer that drives an electron beam through thestron. The environment of the modulator is quite inhostable: there are high voltages with the resulting coronaozone effects and the high voltage is switched quickly withigh-power thyratron, resulting in large electromagneticterference throughout its cabinet and in adjacent equipm
The most important upgrades to the modulator wpower regulation using Sola ferromagnetic resonant poconditioners. Previously both the thyratron and the klystrfilament power would vary during the day and need constadjusting to maintain performance. For the thyratron,nicely isolated, discrete box was custom built by NorthsResearch, rebuilt by FEL Center personnel, and used to dall the connections of the thyratron including the triggpulse. The upgraded power regulation dramatically improvthe operation of the modulator.
Other modifications can be characterized as sound eneering practices. Air cooling was increased to the cabthat houses all the modulator components. The circulatcooling, and filtering of the insulating oil was improved. Thoil blackens over time, presumably due to corona and inquent arcing: filtering the oil keeps it clean and a full ochange is done annually.
rononheroarem
at
inle
mte
ooag
5eel
aan
eaaby
tescinfaanhianb-do
ne
nitfo–
wo
hegeEsheinn
syinati
tinginger.oth
oils aaveen
op-atgndornmsti-ob-at
ofalx-
twoily
ery.IIIfull
ys-tyicalan-
behu-2.5
e tob-ngign-nsn.
plyr-
d
ryportm ismir-g a
3212 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
Another key step towards reaching the desired opetional standards was addressing the technical limitatipresent throughout the building. For example, air conditiing the room that contains the modulator as well as otpower supplies and electronics also turned out to be plematic. A system typically used as a computer roomconditioning system was installed. The system is highlydundant, which has proved invaluable when one of its coponents requires service. In addition, a special chilled wsystem for cooling the linear accelerator~linac! was in-stalled. It was designed and built by Innoventor EngineerInc. and replaced an aging passive system that was unabmaintain the linac temperature especially during the summonths. The design specifications met by the cooling syswere that the linac would be maintained at60.05 F whilerunning the modulator and all the other systems on the cing loop in steady state. It also keeps the temperature steat 60.15 F during interruption of the modulator for as lonas 2 min. It has a variable temperature setpoint of 75–10It also can expel up to 80 kW of heat, although at our presoperating maximum of 30 as opposed to 60 Hz it rarexpels more than 30 kW of heat. The system performsdesigned and has led directly to better spectral stabilitypower stability in the FEL.
The original control system was based on two eight yold 8086-compatible STD-Bus PCs used to perform retime data acquisition and control. EMIs from the nearmodulator periodically rebooted these computers duenoise on the backplane, attributed to aging connectors incomputers and to increased EMI from the modulator. Thtwo 8086 computers were replaced with a dedicated Matosh PowerPC connected over a general purpose interbus ~GPIB! network to several devices, one of which isHP VXI-bus data conditioning and acquisition system. Tdirectly improved operations and increased the amountquality of the diagnostic information from the FEL susystems. The program on the Macintosh was written anmaintained by center personnel and is the first, fastest sware safety check of the subsystems. This Macintosh isworked to another Macintosh runningLABVIEW , which is theoperator interface to the control system.
The optics of the laser cavity and the laser beam traport system were upgraded. The output mirror of the cavis a dielectric coated ZnSe mirror. There are five setsdifferent wavelength bands: 2–3, 2.8–4.2, 4–6, and 6mm, as well as a special mirror with 15% transmission in tbands, at 2.8–3.2 and 5.8–6.8mm. A three-mirror and laterfour-mirror carrousel replaced a single mirror mount in tlaser cavity. Before the upgrade, mirrors were only chanonce or twice a week because of the interruption to Foperation and because of the radiation hazard in accesthe laser cavity midweek. With the four-mirror carrousel tfull wavelength range can be scanned with only modestterruption to lasing while the new mirror is put into positioand aligned.
It is necessary to keep the infrared beam transporttem at rough vacuum pressures because several bandsFEL wavelength range are highly attenuated in air. There7 mirrors between the laser cavity and the laser diagnos
a-s-r
b-ir--
er
gto
erm
l-dy
F.ntysd
rl-
tohee-ce
sd
isft-t-
s-yr9
dLing
-
s-therecs
room, and 14 mirrors between the cavity and the operaroom. With so many mirrors even a few percent of scatterloss per mirror results in large loss of power from the lasThe beam transport system degrades slowly over time bfrom exposure to the FEL and from exposure to air orfrom the vacuum pumps. Mirrors are checked several timeyear to determine if the losses at any given wavelength hincreased. It is of interest that one mirror typically has befound to be the cause of the problem.
There were several important lessons learned duringeration and maintenance of the Mark-III FEL. One is thrunning the FEL 72 h a week nearly every week, allowinsome time Monday morning for routine maintenance astartup, is beneficial for two reasons. First, an operatorengineer likely will be monitoring the control system wheproblems develop, making identification of the probleeasier and quicker. Second, there is a lot of time to invegate, at least via the computer diagnostics, intermittent prlems and plan shutdowns for more invasive diagnosticstimes that are least disruptive.
C. Results
These upgrades substantially improved the operationthe Mark-III FEL in general, and benefited both the medicand nonmedical user communities. The predictability eceeded 95% and serious failures have been limited toweeks per year. As a consequence, this Mark-III FEL readsatisfies the requirements for investigational human surgFurthermore, the spectral and power stability of the Mark-is greatly improved, as well as the ease of scanning therange of infrared wavelengths.
IV. MEDICAL BEAM DELIVERYFOR THE MARK-III FEL
Establishing a stable, reliable, user-friendly delivery stem at the W. M. Keck FEL Center at Vanderbilt Universithat transports the infrared beam from the FEL to the surgsuite approximately 90 m away is a challenging task. Stdards established for medical laser technology need toachieved to satisfy the strict requirements necessary forman surgery. Operating in the wavelength range beyondmm requires unconventional beam delivery techniques duthe broad tunability of the FEL, the strong atmospheric asorption, high peak intensity of the micropulses, and lodistance between the laser source and operating field. Alment, pointing stability, reproducibility, and safety concerof the beam transport system are addressed in the desig
A. Beam alignment
The laser cavity, electron accelerator, and power supof the Mark-III FEL are located in the basement of the foufloor building.20 The fourth floor houses a fully equippesurgical facility, which includes two operating rooms~ORs!~OR1 and OR2!, patient preparation rooms, and a recoveward. Due to atmospheric absorption, the beam transsystem is evacuated and terminates in the OR. The beaeither matched to the entrance aperture of an articulatedror arm using a two-lens telescope or it is refocused usin
single lens into a hollow waveguide based, handheld delivprobe. Figure 3 shows part of the surgical floor plan withlaser contingency room~LCR! and the two ORs. The laseparameters~spectrum, energy, repetition rate, and macro- amicropulse durations! are monitored on an optical table ithe LCR where the beam path can be switched to either
Each OR has two ceiling mounted beam ports~OR1C,OR1W, OR2C, and OR2W!. Beam transport from the LCRto the ORs is achieved by directing the beam upward inLCR onto a mirror at 45°, then horizontally~above the ceil-ing! to the OR. In the ceiling mounted mirror tanks, a 4angled mirror reflects the beam downward into the OR treference point marked by a crosshair. The crosshair incenter of the optical beam path is monitored with a cam~mounted in the ceiling mirror tank! and displayed in theLCR. These crosshairs are the reference points for mounthe articulated mirror arm or launching into a waveguidepossibly a fiber device. The maximum energy per macpulse delivered to the beam ports is about 70 mJ and depon the gain curve of the Mark-III FEL. Consequently, taverage output power is up to 2 W. Mirrors with a diameof 76 mm ~3 in.! ~protected silver coated Si, II-VI, Inc.Saxonburg, PA! are mounted in the turning chambers of tbeam transport system and are equipped with motorizedjustment screws~8301-MRA picomotors, New Focus, San
FIG. 3. Partial floor plan showing the beam path from the laser contingeroom ~LCR! to the operating rooms~OR1 and OR2!.
rye
d
R.
e
aea
ngr-ds
r
d-
Clara, CA! under remote control~a 8732 picomotor multi-axis driver, New Focus, Santa Clara, CA!. Part of the laserbeam (;10%) is split off using a CaF2 beamsplitter~ISP,Tarrytown, NY! and directed through a CaF2 window out ofthe transport system and onto the optical table in the Land is used for diagnostic and alignment purposes. Theput from two pyroelectric quadrant detectors~PQD-220,Molectron Detector, Inc., Portland, OR! is analyzed and usedas feedback signals for the automated alignment system
Figure 4 shows a simplified schematic of the system tdelivers the FEL beam to any one of the four ports in the tORs. The FEL beam enters the fourth floor on the left siM0 and M1 are the main alignment mirrors used to guideFEL beam to the ports near the wall~OR1W and OR2W!.Mirror M1 can be moved out of the beam path by meansa vacuum feed-through lever~MDC Vacuum Products Corp.Hayward, CA!, so that the FEL beam can pass to mirror MMirrors M0 and M2 reflect and align the beam to the ponear the center of the ORs~OR1C and OR2C!. Each of theceiling mounted mirror tanks in the LCR contains two mrors that direct the beam either to OR1 or OR2. The mirr~OR1W, OR2W, OR1C, and OR2C! are drawn elliptically inFig. 4 to represent the fact that the beam is reflected perpdicular to the plane of the page toward the ORs. A phograph of this system in the LCR is shown in Fig. 5.
The alignment procedure coaligns the infrared beamthe FEL with the visible beam of a HeNe laser, whichinitially aligned to the beam ports in the ORs~Fig. 4!. Inbrief, the alignment procedure consists of~1! aligning theHeNe beam to the crosshair in the OR;~2! aligning the twoquad cells, QD1 and QD2~PQD-220, Molectron DetectoInc., Portland, Oregon!, to the split off part of the aimingbeam such that the centroid of the HeNe beam hits the qcell in the center; and~3! using mirrors M0 and M1 to alignthe FEL beam to the center of the quad cells.21 A computercode was written inLABVIEW ~National Instruments, AustinTX! that uses the signals of the quadrant detectors as f
y
FIG. 4. Schematic of the beam switching and alignment system on the optical table in the laser contingency room.
lgndrahbysewt
l oeN0p
fu
5°er
powoualthir
ohigedgir
eoth
elarasut-desept-ys-
beheldgicalx-
ril-ashile
-de-
ery.ed
hasas
ass
abar
rrorllow
3214 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
back to align the mirrors~M0 and M1!, thereby controllingthe picomotors on the mirror mounts as actuators. The arithm aligns the FEL beam with M0 to the center of QD1 athat with M1 to the center of QD2 in an iterative loop. Aftethe infrared FEL beam is aligned to the center of the beport, beam splitter BS1 can be moved out of the beam patmaximize the power of the FEL to the OR. This is donemeans of a vacuum feed-through lever. Considering thesitivity of the quadrant detectors, parallelism of the windoand the beam splitters, accuracy of the mirror mounts andpiezoelectric screws, and proportional feedback controthe computer code, the current setup can coalign the Hbeam and the infrared beam within a divergence of aboutmrad, that is, the center of the beams can be displaced u2 mm over the 20 m path length between the LCR andthest OR port.
In the ceiling mounted mirror tanks in the OR, a 4angled mirror reflects the IR beam down into the OR whit transmits through a 10 cm diam, 1 cm thick BaF2 windowthat serves as the final window in the evacuated transsystem. The beam ports in each of the ORs are equippedprecision optic mounts to adjust the waveguide or fiber cpling device or the articulated mirror arm to its final optimposition. Figure 6 shows the delivery handheld probe andarticulated mirror arm connected to the ceiling mounted mror tanks in the ORs, and gives a good overall impressionthe actual design. The optomechanical mounts and a phdiode connected to a pulse counting unit are inside the wenclosure hanging from the ceiling. This enclosure is purwith nitrogen in order to avoid atmospheric losses. An adtional nitrogen port is provided as a purge line for the surcal beam delivery device. The surgeon controls a shuttethe beam path by a foot switch~pedal!. This shutter defaultsto its closed position.
B. Beam stability
We investigated the stability of the delivery system ovthe 90 m path length; in particular, we were concerned abvariations in temperature in the building. We measured
FIG. 5. Photograph of the laser contingency room showing the optical twith the capability for the beam switching and alignment. QD1 and QD2the quadrant detectors outside the evacuated beam transport system.
o-
mto
n-shefe
.1to
r-
e
rtith-
e-ofto-ted
i--in
rute
beam stability in the OR over the course of 1 h, at a timtypical for a surgical procedure. The resolution for angudrift of the beam was 0.06 mrad, i.e., the beam drift w,5% of the beam diameter while the temporal energy oput varied'5%. However, if the FEL needs to be retuneduring a procedure, the infrared beam typically becommisaligned and the transmission of energy can be unaccably reduced. In this case, realignment with the current stem can be accomplished within 5 min.
C. Surgical delivery device
The final stage of the beam delivery system needs toa flexible, maneuverable device that terminates in a handprobe amenable to the surgeon to use as a precision surinstrument. The key specifications for this device are fleibility, maneuverability, ruggedness, and ability to be steized. In addition, depending on the clinical application,much as 30 mJ per macropulse needs to be delivered wfocusing onto a spot size of 300mm diameter. Due to wavelength range and peak irradiance constraints, fiber opticlivery is at the present time not feasible.22 We have imple-mented two alternative strategies for surgical beam delivFirst, a modified articulated arm was outfitted with infrarmirrors and terminated with a focusing lens (f 5125 mm).This proven technology gives six degrees of freedom,relatively low losses, and is robust. We have deliveredmuch as 50 mJ/macropulse to tissue in a 300mm spot sizewith this articulated arm. The major disadvantage is the m
lee
FIG. 6. Photograph of the two beam ports in OR2: The articulated miarm is mounted on the left and the handheld probe based on the howave guide is mounted on the right.
~even though it is counterweighted! and relatively large sizeNevertheless, this was the delivery system of choice for nrosurgery. As an alternative, delivery systems based onlow waveguides~HWGs! have been developed.23 The nu-merical aperture~NA! equivalent of the waveguide wameasured to be 0.06 atl56.4mm ~which corresponds to anacceptance angle of 3.4°).22,24–26 The final design of theprobe22,24 has a sealed tip, contains a CaF2 microlens to fo-cus the beam to a spot size of;200mm, and has a workingdistance~focal length! of approximately 3 mm. The limita-tions of the HWG-based delivery device are significalosses (.50%), variable losses depending on bending ofwaveguide, and the potential of damage to the interior wof the waveguide. Nevertheless, this device has been usophthalmic applications where delivery of only 3 mJ pmacropulse was sufficient. Investigations of stretching ofmicropulse duration due to transmission through a HWGdescribed elsewhere.22,24
D. Loss analysis
The total energy loss in the beam delivery system issum of the losses of the optical elements and in additionabsorption in the atmosphere for certain infrared walengths. There are seven mirrors located between the outpling mirror of the Mark-III FEL and the Laser Control room~located on the second floor!. Four more mirrors direct thebeam onto the optical table in the fourth floor LCR. ThuM0 is the 11th mirror in the beam transport system. Frthere three more mirrors are used to direct the beam to anthe OR output ports. While single elements have low los~typically 98%–99% reflectance, 92% transmission in a leor window!, the cumulative loss totals 30%–35% from thcontrol room to the point of launching into the delivery dvice ~hollow waveguide or articulated arm!. The optical ele-ments that contribute to this loss are seven mirrors, one B2
window ~the output window and the seal of the vacuum stem!, and one lens (f 5300 mm, in the case of the waveguide! or two lenses (f 5300 and 75 mm, in the case of thbeam reducing telescope needed to couple to the articuarm!. In addition, the delivery device itself causes significalosses. Coupling losses for the hollow waveguide are 3040% and attenuation is 1dB/m, for a typical net lossgreater than 50%. In contrast, losses in the articulateddue to the seven mirrors and the final focusing lens toapproximately 20%.
E. Power stability
The macropulse energy of the Mark-III FEL is wavlength dependent, with a maximum at approximately 6mmthat falls off at wavelengths below 2.5 or above 8.5mm.6,20
The output near 6mm often exceeds the requirement fhuman surgery, necessitating the capability for beam atteation. A double Brewster plate polarizer made of Zn~PAZ30mm-AC, II-VI Inc., Saxonburg, PA! is installed in-side the vacuum beam line, upstream of the first mirror~M0!in the LCR. This polarizer is mounted to a motorized rotional picomotor stage~8401, New Focus, Santa Clara, CA!and interfaced to a multistage controller~the 8732 picomotor
u-l-
tellin
ree
ee-u-
,
ofss
F-
tedt–fml
u-e
-
multiaxis driver, New Focus, Santa Clara, CA!, as are all thecontrols for the picomotor-driven mirrors. Since the FEbeam is linearly polarized, rotating the polarizer reducespower delivered to the target~polarization extinction is1:3000!. The position of the polarizer was chosen such thacan be used regardless of the surgical output port thaselected and is in front of the final alignment in the LCRotation of the polarizer results in negligible misalignmeof 0.01 mrad.
F. Results
The design and implementation of the medical beamlivery system for the Keck FEL Center met the safety aperformance standards required for human surgery. Howefurther improvements of the resolution and reproducibilitythe alignment system will be necessary for clinical use oregular basis.
V. NEUROSURGICAL APPLICATIONSOF THE MARK-III FEL
Previous investigations of neural tissues demonstrathe preferred ablation properties of the Mark-III FEL whetuned to wavelengths near 6.45mm.8 Survival studies of FELincisions of canine brain tissue27 as well as of survival ratand isolated rat brains8 showed a marked reduction of collaeral thermal injury using FEL wavelengths near 6.45mm. Inthe case of brain tissue, the damage to adjacent tissuecollateral thermal injury, measured by histology was appromately one cell width deep to the laser incision and typicaundetectable on the sides of the incision. Because ofpromise shown by this series of studies, a protocol was idtified for human surgery which met with the approval of bothe FDA’s Investigational Device Exemption~IDE! andVanderbilt’s Investigational Review Board~IRB!.
The protocol called specifically for the partial FEL excsion of an extra-axial brain tumor.28 In other words, a smallexternal portion of a tumor would be laser excised, and ththe rest of the tumor would be removed using traditionmethods. Patients with anextra-axial tumor, or tumor thatoriginated inside the skull but outside the brain, were chobecause no normal brain tissue would have to be traversehave access to the tumor. In this way, the number of posssources of brain injury for the procedure is reduced and tinterpretation of any poor neurologic outcomes postopetively is greatly simplified. Extra-axial tumors, such as a spcific type called meningiomas, also are characteristicanearly always benign. Therefore, long term survival of tpatient is much more likely, should long term side effecbecome an issue. Finally, such tumors can usually bemoveden bloc, or intact, so that the tissue adjacent to tablative volume will likely be spared from being lost in thremoval of the remaining tumor mass and can thus be sied histologically without having to fit pieces of a tumopuzzle together. While these tumors in general are ultimanot good candidates for laser resection because of their tcal ease of resection with traditional nonlaser based meththey proved to have many qualities that are desirable forinitial study.
all
nEfa-ele
thetoroE
aif
t iLpull
igy
Tht
thitomanthum
ht
theor
ed,hetheiga-erhe
ible
ofliv-theere
o-
Byideieceorl-
po-lat-rre-
ofro-
m
ofm.
3216 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
A. Human neurosurgery
The first patient had a suspected meningioma of at le3 cm in size, as shown in Fig. 7, and was neurologicastable prior to the procedure. The patient was operated oan experienced neurosurgical team in Vanderbilt’s Keck FCenter. While this was the first human operation in thiscility utilizing the Mark-III FEL, the facility had been performing both conventional laser and nonlaser human surg~primarily orthopedic procedures! on a daily basis for severamonths prior to this operation. There are extraordinary nefor a neurosurgical operation of this complexity. It becamclear early on in the design of these operating roomsgreat care was required to minimize the likelihood of forgting a seemingly insignificant device, medicine, supply,part. Thus a track record was established of successful,tine cases using these operating rooms prior to the first Fcase.
The patient was brought to a holding area from the mhospital, where anesthesiologists prepared the patientgeneral anesthesia. The patient was prepped and broughone of the two operating rooms fitted for use of the FEOnce under anesthesia, the patient was positioned in a suposition with a shoulder roll under her right shoulder, thallowing her head to be easily positioned nearly horizonta~Fig. 8!.
A standard neurosurgical approach was used in the rtemporal area, guided by a Pickar® neuronavigational stem which utilized a computer topography~CT! scan thatshowed contrast obtained the day prior to the procedure.system enables a computer in the operating room to showlocation and position of our instruments with respect topatient’s anatomic features by viewing a computer monloaded with the preoperative CT images. This is accoplished using a system of infrared emitters on the patientour instruments, and a series of infrared cameras withinroom all of which feed back to the computer. Using a neronavigational system thus allowed the creation of a coparatively small 6 cm curvilinear incision over the rig
FIG. 7. Preoperative MRI scan~axial T1 weighting with gadolinium con-trast!. The arrow points to the tumor and its compression of the right teporal lobe.
stybyL-
ry
dseat-ru-L
nornto.inesy
hts-
isheer-de--
temple to expose the skull over the tumor, rather thantypically large skin flap one uses just to ensure that the tumis within the exposed area. A 5 cm bone flap was raisexposing the underlying dura, which lines the inside of tskull. The tumor, attached to the deep, or brain side, ofdura, was easily located using the wand of the neuronavtional system. A 1 cm cuff of dura was cut around the bordof the tumor and then folded back to expose the tumor. Tsurrounding normal brain was protected from any possstray laser energy by moistened cottonoids~Fig. 9!.
Once surgically exposed, about 1500 macropulses6.45 mm radiation averaging 32 mJ/macropulse were deered through an articulated arm to the outer surface oftumor. The articulated arm terminates in a hand piece, wha calcium fluoride lens with a focal length of 12.5 cm fcused the FEL beam to a spot size of 310620mm. Ahelium–neon pilot beam was used to guide the excision.practicing one final time with some sterile paper on a stable, the surgeon was able to then position the hand pwith the visible pilot beam to keep the focus on the tumsurface, controlling FEL-beam delivery with a foot-pedaactuated beam block. In the first operation, total FEL exsure administered in multiple passes took about 50 s, abing approximately a cubic centimeter ‘‘divot’’ from the outesurface of the tumor. The remainder of the tumor wassecteden bloc~Fig. 10!.
Once removed, the tumor was divided and a portionthe tumor that did not include the defect was sent for neu
-
FIG. 8. Position of the patient prior to incision. The arrow shows locationthe tumor~circled ‘‘T’’ ! based on data from the neuronavigational syste
FIG. 9. ~Color! Exposed tumor just prior to lasing.
pathologic examination to confirm the diagnosis. Themainder was processed, stained, and mounted on slidesinspected under the microscope. The width and depth ofdivot were measured for each of the sections. The adjatissue was then inspected at a minimum of five locationscoagulation necrosis was measured at approximately th5:30, 6:30, and 8 o’clock position of the section~shown byblack arrows, Fig. 11!.
B. Postsurgical assessment
The laser performed extremely well and was indistguishable from previous animal experiments. The tumorsue was ablated quickly and effortlessly. The surgeon notsignificant subjective advantage of this type of resection wrespect to effort and speed in comparison to other commmethods of resection used for meningiomas, specificallytrasonic aspiration, bovie loop resection, or mechanicalmoval with scalpel or forceps. Each macropulse ablateportion of tumor approximately 0.5 mm wide and 0.07 mdeep and the articulated arm was easily and predictablytrolled with the foot pedal. An estimate of the rate of ablatiwas approximately 1.8 mm3 per second, depending on hanspeed. There were no limitations in terms of angle of ex
FIG. 10. ~Color! Excised tumor. The arrow points to a vaporized defec
FIG. 11. ~Color! 10 mm section of meningioma, stained with H&E. Ablate‘‘divot’’ indicated by black arrows, at approximate sites for measuring psible coagulation necrosis. The white arrow points to the area of the tiprocessing artifact.
-ndentd4,
--a
hnl--a
n-
-
sure of the laser to the tumor. There was no char. Therea smoke evacuation apparatus for safety, but no smokevisualized. On one subsequent patient, the meningiomaclearly calcified preoperatively, as evidenced on a CT sand by its texture at the time of surgery. The ablation rwas noted to be much slower in that case and calcific lesmay be more efficiently resected at another wavelength.
The gross histologic examination of the slides~Figs. 11and 12! showed remarkably sharp, square edges at the latissue interface with minimal hemorrhagic deposition~show-ing as small magenta colored blobs! at the base and sides othe lesion. Adjacent tissue injury typically ranged from udetectable to one cell width and at no time was found togreater than three cell widths. This finding is similar to rsults found in previous animal experiments in our labs. Ulike CO2 laser lesions, there was no char on any of tsections.29 There was also no discernable ‘‘pale zone’’ likthat described around lesions made by other surgical lase29
The presence of well defined nuclear morphology, typicalost in tissue coagulation necrosis from collateral therminjury, is evident next to the very edge cells adjoining tdivot throughout, suggesting that much of the adjacent tisinjury could be attributed to nonthermal processes. Coagtion necrosis, or heat damaged tissue next to the lesion,judged as minimal to none. Other such processes couldclude mechanical injury due to the pulsed nature of the lasimple perioperative tissue hemorrhage, as well as tissuecessing artifact.
From a clinical standpoint, there were no complicatiofrom any of the first three patients involved in this study. TFEL has been shown previously not to have significantmostatic properties,27 however there was no significant hemorrhaging during laser ablation. The neuropathogy of eachthe tumors resected was consistent with meningioma, asdicted by preoperative scan characteristics. Postoperativthe patients woke up quickly and continued to have an exthat was unchanged neurologically. A postoperative magnresonance imaging~MRI! scan one week postop showed nresidual tumor, nor any evidence of ‘‘punch through’’ of th
-ue
FIG. 12. ~Color! Closeup of the 5:30 o’clock position in Fig. 5~magnified203). Note the squared off appearance of what looks to be a divot duesingle pass of the FEL beam. Also note the presence of the intact numorphology ~small purple dots! near the edge of the divot, with cellulamorphology lost only in cells adjacent to laser exposure that appears tconsistent more with cellular edema rather than with true coagulation nesis traditionally found in collateral thermal injury.
lce
aityinicau
ud
dr
abten
mreouorro
o-tioatrthithhaa
san
roavei
n-izedbera-nceflec-ra-thisf tothe
aandctionaton-g
Lbiltumedsedry
the150
hen-nd-
fore.id-ith a
lises
bye
.bore
forntoyent
-lo
3218 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
laser ~Fig. 13!. Follow-up examinations found no clinicaevidence of adjacent or distant brain injury from the produre.
The first human application of a free-electron laser wsafe and successful, with the initial results consistent wprevious animal data. With respect to ease of use, efficactumor resection, and minimization of collateral thermaljury, the FEL was judged to be superb in this neurosurgapplication and compares very favorably to conventional sgical lasers. These findings support continuing clinical sties with the FEL and suggest that the FEL tuned to 6.45mmhas potential as a clinical surgical tool with significant avantages over conventional surgical lasers for use in thesection of certain tumors in the brain.
C. Future applications
Computer driven mirrors have already been well estlished both in science and industry to give very intricacontrol of lasers in making even very complex patterns, awith much more speed and accuracy than control by a huhand. Using a laser with such control and also the appaaccuracy for removing tissue demonstrated above, one ceasily imagine an automated system for resecting tumsuch as that already being done with vision correction pcedures.
While a surgeon would still likely need to provide expsure, or a pathway to the tumor, an image-guided reseccertainly has the potential to be faster and more accurSeveral image guided systems have reported submillimeaccuracy, including the one used in this study to localizetumor. Most would agree this compares very favorably whuman knowledge and intuition-based navigation. Wwould be needed that does not yet exist in proper form isaccurate system of real-time image feedback, so that thetem can be ‘‘aware’’ of what tissue has been resectedthus the anatomy of the changing target.
There are a number of ways one could envision intducing a feedback loop. A few medical centers already hintraoperative CT or MRI scanners constructed within th
FIG. 13. One week postoperative MRI scan of the first patient~axial T1weighting with gadolinium contrast!. The arrow points to the previous location of the tumor. There is no abnormal enhancement and the temporalhas filled the tumor void.
-
shof-l
r--
-e-
-
danntlds,-
nte.ice
tn
ys-d
-er
operating room facilities. A rescan would likely be time cosuming, but accurate. Ultrasound has already been utilfor navigation systems and many tumors would likelygood candidates for that modality. Interpretation of the ultsound information can be a bit fuzzy at times, however, siimages can be less than crisp depending on the sound retion qualities of the tissue. Attenuated total reflectance infred spectroscopy also shows diagnostic potential fortask, and thus we may be able to use the infrared itselhelp gather the information needed for reconstructingchanging environment.
While certainly much testing would be required, andwell-trained surgeon would be needed for both exposuresupervision of all such cases, the ease and speed of resedemonstrated by the Mark-III FEL in this study suggest than automated system for resecting tumors is not only cceivable, but could likely be accomplished with existintechnology within five years.
VI. OPHTHALMIC APPLICATIONSOF THE MARK-III FEL
A. Delivery system
Ophthalmic surgical procedures using the Mark-III FEhave been performed at the Keck FEL Center at VanderUniversity. The infrared beam is transmitted under vacufrom the FEL vault to the surgical suite, where it is directinto a nitrogen-purged gantry. Then the FEL beam is focuinto a hollow-glass waveguide to obtain a flexible delivesystem in the surgical field~Fig. 14!.30 An adjustable dia-phragm is placed in front of the focusing lens to regulatetransmission of the single-mode, Gaussian FEL beam. Amm CaF2 lens then focuses the beam into a 550mm spot.The beam then passes through a 500mm diam pinhole thatprotects the walls of the waveguide during alignment. T530 mm inner diam waveguide is attached with a SMA conector and protected with Teflon tubing. The surgical hapiece ends in a 20 gauge thin-wall cannula. A CaF2 lens witha focal length of 2 mm is placed at the tip of the cannulafinal focusing and to protect the waveguide from moistur
The hollow-glass waveguides are fabricated by a liquphase deposition technique that coats the inner surface wsilver film and then with a AgI film.31 The thickness of theAgI coating will alter the wavelength optimization. Marcatiand Schmeltzer theoretically accounted for the low losassociated with these waveguides in 1964.32 Bending lossesfor hollow-glass waveguides were calculated in 1990Miyagi and Karasawa33 where the key dependencies of thattenuation coefficienta were
a}l2
a3 and a}1
R, ~1!
wherea is the inner diameter andR is the bending radiusThus there are losses with bending and a decrease insize.
For our surgical purposes, waveguides optimizedamide bands I, II, or III were selected and assembled iprobes. At amide II~6.45 mm!, losses were determined bmeasuring input and output energies through two differ
FIG. 14. Schematic of the FEL hollow-glass waveguide delivery system. The FEL beam passes through an adjustable diaphragm to attenuate thethrough a 150 mm focal length lens and a 500mm diam pinhole to couple the beam to the hollow waveguide. The waveguide is within a surgical probis protected with a CaF2 lens at the tip.~Reproduced with permission from the Optical Society of America; see Ref. 30.!
rv
ul-
bf
o
uloe
m-cut-
veg
suree-nes-ngednrve
es-sThe
e-
Nol
waveguide lengths, as well as by a 360° 25 cm radius cuThe coupling loss was 160.3 dB (2065%), thetransmis-sion loss was 2.160.15 dB/m~16.5%!, and the loss in cur-vature was 0.760.25 dB (1565%).30
A surgical probe was sterilized with ethylene oxide mtiple times to permit sterile surgery on animals~Fig. 15! overa year without significant degradation. A waveguide prodelivering a 6.45mm infrared beam with a fluence o2.8 J/cm2 and a spot size of 300mm in diameter were used tablate cadaver retina~Fig. 16!.30
B. FEL optic nerve sheath fenestration
Optic nerve sheath fenestration has been successfpreventing vision loss in patients with pseudotumcerebri.34 This is a disease where abnormally elevated pr
FIG. 15. Surgeons using a hollow-glass waveguide FEL surgical probperform an animal experiment.~Reproduced with permission from the Optical Society of America; see Ref. 30.!
e.
e
inrs-
sure of the cerebral spinal fluid surrounding the nerve copresses and damages the fibers. This procedure involvesting a window in the thick coverings surrounding the nertransmitting vision to the brain. It is technically challenginto expose and cut a window in the sheath to relieve preswithout damaging the underlying optic nerve. This wavguide system was used to perform optic nerve sheath fetration in rabbits, promising a technically less challengiprotocol.35 Routine histological tissue data analysis showthat the 6.45mm FEL beam at 10 Hz macropulse repetitiorate and 2 mJ/macropulse were able to cut the optic nesheath without cutting the underlying nerve.36
Anesthetized rabbits received optic nerve sheath fentrations with the FEL or with a knife. The conjunctiva waopened and the superior rectus muscle was disinserted.
toFIG. 16. Histologic results showing the ablated incision in the retina.collateral damage was found.~Reproduced with permission from the OpticaSociety of America; see Ref. 30.!
Ave
0
-trdn
ith
ie2
arcup
d
ithn
teof
lingageap-tionJ.
u-riorck-y.
re-
en-to
ma
d toad-res
es-ivedlat-
L
myyf
loudy.t
3220 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
optic nerve could be observed with extensive retraction.mm diam window was produced by incising the optic nersheath either with a knife or with the FEL beam~6.45 mm,10 Hz macropulse repetition rate,,2.5 mJ/macropulse, 30mm spot size, 5ms macropulse! using a 530mm diam wave-guide probe@Fig. 17~A!#. Then using a small hook we carefully removed the dura and arachnoid to create the fenestion window @Fig. 17~B!#. The superior rectus muscle anconjunctiva were repaired. The rabbits survived for omonth or were sacrificed immediately following surgery wtheir optic nerves prepared for histologic analysis.35
Aiming the FEL probe was found to be technically easand relatively efficient at cutting the circumference of themm diam circle in the small space between tissues compto positioning the knife. One or on rare occasions two cirlar treatments with the FEL lasing an average of 2.0 mJmacropulse were adequate to incise the dura.35
Both 6.45mm FEL incisions at,2.5 mJ/macropulse an
FIG. 17. ~Color! ~A! Sterile hollow waveguide probe that delivers FEenergy to the optic nerve sheath.~B! Window over the optic nerve is visibleafter removal of the incised sheath~shown by the arrow!. ~Reproduced withpermission from Ref. 35.!
2
a-
e
r
ed-er
knife incisions showed a lack of optic nerve damage whematoxylin and eosin~H&E! staining. Near the fenestratiosite, glial fibrillary acidic protein~GFAP! ~a histologicalmarker of astrocyte activation! was increased within a fewhours after incision with the FEL or the knife to demonstrathe sensitivity of these cells to manipulation. Hypertrophythese glial cells was also evident after one month of heaequally in both treatment groups. Extensive nerve damwas observed when a macropulse energy of 7.5 mJ wasplied to one nerve sheath, but optic nerve sheath fenestraappeared safe with macropulse energies less than 2.5 m35
Additional studies are ongoing.
C. Endoscopic FEL goniotomy
Goniotomy is a surgical treatment for infantile glacoma. In goniotomy, a fine needle is inserted into the antechamber through a peripheral corneal incision to cut thiened trabecular beams to allow the iris to move posteriorl37
This may reduce intraocular pressure by decreasing thesistance to aqueous outflow. A 250mm inner diam wave-guide system was combined with a 0.8 mm diam oculardoscope to perform FEL laser goniotomy and to compareneedle goniotomy in anesthetized congenital glaucorabbits.38
Goniotomy was performed with either the [email protected]~A!# or the FEL@Fig. 18~B!# at 6.45mm, 30 Hz macro-pulse repetition rate, and 2.2–3.5 mJ/macropulse couplean endoscope. The rabbits’ corneal edema preventedequate visualization of the anterior chamber angle structuthrough the cornea@Figs. 18~A! and 18~B!#.38 The image ofthe angle was viewed on a video monitor@Figs. 18~C!–18~E!# as the angle was incised 100° – 120°. Intraocular prsures were measured postoperatively. The animals survthree weeks and then underwent goniotomy on the contra
FIG. 18. ~A! 0.8 mm diam endoscope coaxially coupled to a goniotoneedle placed in the anterior chamber.~B! 0.8 mm diam endoscope coaxiallcoupled to a 250mm diam hollow waveguide for intraocular delivery oFEL energy placed in the anterior chamber.~C!–~E! Anterior chamber struc-tures viewed through the endoscope since the corneas are extremely c~C! The needle tip is placed into the angle.~D! The laser probe is aimed athe trabecular beams.~E! An area of incised angle is visible.~Reproducedwith permission from Ref. 38.!
eral eye immediately prior to euthanasia. Both eyes wpreserved for histologic evaluation.38
This FEL surgery successfully lowered the postoperaintraocular pressure for three weeks in the rabbits andcomparable histologically to endoscopic goniotomy pformed with the needle. No collateral thermal damage wobserved.38
D. Future prospects
The Mark-III FEL beam delivered through a hollowglass waveguide has the potential for surgical applicationophthalmology. It has been delivered successfully to smsurgical targets in and around the eye. A clinical comparitrial of optic nerve sheath fenestration with the FEL or scsors on blind eyes prior to enucleation is ongoing. In ordebecome clinically useful and acceptable for surgical produres, smaller more economical lasers that mimic theoperating characteristics of the Mark-III FEL must be devoped.
VII. FEL BEAM DELIVERY VIA EVANESCENT WAVECOUPLING AT OPTIC–TISSUE INTERFACES
Historically, precise laser surgery has targeted walengths where the optical absorption in the tissue is high.micron-scale surgical precision using normally incident lapulses, high tissue absorption coefficients in the ultraviobelow 220 nm,39 and the infrared, near 2.94mm,40 are typi-cally targeted with conventional lasers. However, microscale precision can be obtained using evanescent waeven at wavelengths where optical absorption is weak. Usevanescent optical waves, laser energy can be confinedlayer less than one wavelength thick at the surface of hrefractive-index devices.41,42 Optical energy outside of thedevice is present only within a few micrometers of the dvice’s surface, rather than freely propagating as a la‘‘beam’’ into the tissue. The Mark-III FEL is ideal for achieving efficient ablation rates because of its wavelength tunaity, short pulse duration, and high average power. For ecient ablation of tissue with minimal thermal damage, shpulse durations that achieve thermal confinement43,44 on thismicron scale is necessary. Each short laser pulse of sufficincident energy removes a layer of thickness approximatedand thermally denatures a layer of several timesd.45 Theoptical penetration depthd is important not only for deter-mining the amount of tissue removed and the residual thmal damage, but also for choosing the optimal laser walength and pulse duration. For example, for a 1mm depth ofpenetration the estimated ideal pulse has a duration of a1 ms and an energy of at least tens of millijoules. A Mark-FEL is well suited to these requirements. The pulse durais variable over the desired range and the wavelength catuned through the midinfrared tissue absorption bandsstudy the deposition of energy, cavitation~vaporization ofwater!, and tissue ablation arising from evanescent waveteractions.
Maxwell’s wave equations can be applied to describeevanescent optical fields in water~a good approximation forwet tissues! for all angles of incidence and polarizations
re
eas-s
inlln-o-y-
-orrt,
-es,g
o a-
-er
il--
rt
nt
r-e-
ut
nbeto
-
e
interfaces with rugged visible-infrared optical materials suas silica, sapphire, zinc sulfide, silicon, and germaniuBased on this analysis, the useful evanescent field depththe order of 0.1–2mm.42 Here in Sec. VI we will describehow evanescent optical waves are generated at optic–tiinterfaces, and how these evanescent waves can be useagnostically to sense, monitor, and measure the tissue,therapeutically to thermally injure or ablate the tissue.
A. Evanescent optical waves
Electromagnetic radiation is totally reflected from an iterface defined by a high-refractive-index mediumn1 and alow-refractive-index mediumn2(n1.n2) when the angle ofincidence exceeds the critical angle. The critical angleuc fortotal internal reflection is defined by Snell’s law, sinuc
5n2 /n1. The boundary conditions require that the electfield, and hence the energy, be present in a layer somewless than one wavelength thick on then2 side of the inter-face. The waves in this layer are called evanescent wabecause they decay rapidly to zero away from the interfaThe plane wave evanescent electric field amplitude intransparent medium is given by46
Ee~x,y,t !5Ee exp@ i ~kxx2vt !#exp~2gkyy!, ~2!
whereg5(n12 sin2 ui2n2
2)1/2, andky andkx are the wave vec-tors in the planes perpendicular and parallel to the interfaSince the wave fronts or surfaces of constant phase~parallelto theyz plane! are perpendicular to the surfaces of constamplitude~parallel to thexz plane!, Eq. ~2! is a inhomoge-neous wave. The power~irradiance! is proportional toEe
2 exp(22gkyy). The penetration depthde of the evanescenwave in a absorbing external medium is
de51/~2gky!5l/@4p~n12 sin2 u i2n2
2!1/2#. ~3!
The amplitude of the evanescent wave decays rapidly iny direction and becomes negligible at a distance of onlfew wavelengths of light. When the lower refractive indemedium has absorption, the indexn2 is replaced by the magnitude of the complex refractive index,un2u5n2,r2 in2,i ,where n2,i is the absorption index and is defined asn2,i
5lma/4p. This introduces the absorption coefficient of thexternal medium~e.g., tissue! into the solution of Eq.~2! anddescribes the loss of ‘‘total’’ internal reflection due to absotion within the evanescent wave. The reflectance atoptic–tissue interfaceR',i can be written according toFresnel relations where' andi are the polarizations perpendicular and parallel to the plane of incidence, respectiveThe fraction of absorbed incident energy in water~mimick-ing wet tissue absorption in the infrared spectrum! is givenby (12R',i). The required incident energy is obtained bmultiplying the latent heat of vaporization of water by thpenetration depthde and the laser beam area, and dividingthe absorbed fraction@2500 J/cm33de3(pv0
2)/2/12R',i#,and may be a factor of 8 less if the partial vaporization moholds forEv'330 J/cm3.44
Figure 19 shows evanescent optical waves at an interbetween sapphire and water. The leftmost part of each cucorresponds to the critical angle for that wavelength, for
4
3222 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
FIG. 19. Evanescent optical waves at an interface between sapphire and water.~A! The penetration depth of the evanescent wave and~B! the incident energyrequired for the vaporization of water by laser energy for perpendicular polarization as a function of the incident angle for wavelengths from 2 tomm.
fio
a-
visuit-siit-
ene-
.f
rse
troesen
for
s forath-,
if-ag-
fer-theter-ivethee
are
e
se
theab-n-
stance,uc;48° at 2.1mm. Note the variation in the order omagnitude in the evanescent wave’s depth of penetratThe laser beam area calculation assumes a 500mm Gaussianbeam waist,v0 . The input energy required to induce abltion in water at a sapphire–water interface using a 500mmGaussian beam waist is less than about 30 mJ for all walengths except 2.1mm, where the input energy requiredroughly 100 mJ. These are modest energy per pulse reqments for a Mark-III FEL. The practicality of evanescenwave-driven tissue ablation is apparent from this analyEvanescent wave ablation occurs at modest power dens(;105 W/cm2), well below the threshold for optical breakdown (;109 W/cm2).42
Figure 20 shows tissue ablation using normally incid(0°) and evanescent (65°) optical waves at a sapphirporcine aorta interface with Mark-III FEL light at a wavelength of 3.24mm (' polarization!. The images are toluidineblue-stained histology sections@bar5100mm in Fig. 20~B!and bar520mm, ablation to the left of the arrow in Fig20~B!#. The image in Fig. 20~A! shows an ablation depth oabout 500mm, and that in Fig. 20~B! shows an ablationdepth of about 4mm. Also visible are the black elastic layeand the smooth gray muscle cell layers that underlie thedothelial cells at the surface.
Evanescent waves have long been used in ATR speccopy in the infrared~IR!,47 in which the evanescent wavfield is used to measure absorption spectra at surfacecontact with high-refractive-index crystals. Evanesc
n.
e-
re-
s.ies
t–
n-
s-
int
waves at the boundary of fiber optics have also been usedspectroscopy and diagnostic applications.48 Now supposethat we design a catheter that launches evanescent waveboth diagnostic and therapeutic procedures. The same ceter can deliver low level diagnostic light, say, from FTIRand therapeutic light from an IR FEL, for instance.
Our preliminary work on the spectral signatures of dferent tissues in the IR shows promise for diagnostic iming. Figure 21~A! shows IR reflectance spectraR of porcineaorta and fat tissue taken with an ATR cell from 2–10mm.Note the differences in the spectra near 3.4, 5.7, and 8.5mm.These spectra give a nice qualitative picture of tissue difences as a function of the wavelength. In order to modeltherapeutic interaction of the evanescent wave at the inface, though, it would be much better if we had quantitatdata for the optical properties of the tissue, for instance,complex refractive indexn2 . This can be done by taking threflectance data setR from the ATR FTIR, input it into aKramers–Kronig algorithm to obtain the phase, and compthe Fresnel reflection at the interface to calculaten2,r andn2,i
parts of the complex refractive indexn2 .49 Figures 21~B!and 21~C! show our calculation of the complex refractivindex of the tissues from Fig. 21~A!. This now allows moreprecise modeling of the light–tissue interaction at theoptic–tissue interfaces.
It is also possible to launch an optical probe beam atinterface to sense and monitor the dynamics of a tissuelation process.50 This probe is capable of measuring the e
t at a
FIG. 20. Tissue ablation using~A! normal incident (0°) and~B! evanscent (65°) optical waves at a sapphire–porcine aorta interface with FEL lighwavelength of 3.24mm with perpendicular polarization.
ergy deposition at a dielectric–tissue interface. The transrise in temperature of the laser-heated layer at the intercan be measured using this calibrated probe and the onsablation can also be detected.
FIG. 21. ~Color! ~A! IR reflectance spectra of porcine aorta and fat fromto 10mm. Note the differences in the spectra near 3.4, 5.7, and 8.5mm. Thecalculated~B! real and~C! imaginary parts of the complex refractive inden2 .
ntcet of
B. Future prospects
We are presently investigating the uses of these evacent optical waves for applications in cardiology, neurosgery, and laparoscopy. The IR catheter we are developwould allow spectroscopic feedback at the time of surgerybetter determine the margin between healthy and diseatissue. In cardiology, identification and modification of aterosclerotic plaque using IR evanescent light delivethrough a catheter51 may allow selective ablation of plaque athe unique wavelengths of fatty tissue absorption shownFig. 21. Preliminary data for neural tissue hold promiseidentification of brain tumors, and the precise, controlledlivery of evanescent waves for ablation may allow betresection of these tumors with less damage to the surrouing healthy neural tissue. The ability to dissect throughfatty tissue surrounding many of the organs of the abdomwithout transecting the blood vessels would be a majorvance in laparoscopy. The unique spectral signatures of thfatty tissues in the IR may make controlled, selective abtion of fatty tissue in the brain, blood vessels, abdomen,spinal cord possible.
In contact geometry, for instance, an optic–tissue intface, evanescent waves have a unique capability in diagnand therapy, where precision and control of light is imptant. Laser-generated evanescent waves can achievetremely precise superficial ablation of tissue. Compared wnormal-incidence exposure, the depth of ablation is limiby the evanescent wave penetration depth, independenthe number of laser pulses, and can be orders of magniless than the normal-incidence ablation depth. The practimplication is that high-precision, endoscopic laser surgidevices can now be realized, with the added advantagcontrol of the laser energy. Unlike free-beam laser surgonly tissue in contact with the optical interface of an evancent wave device is ablated, thereby allowing safe infrabeam delivery in the operating room. These tools may allthe diagnosis~spectroscopy! and therapy~ablation! to be per-formed in one catheter.
VIII. MICROBEAM TO INVESTIGATE TISSUEDYNAMICS
The application of molecular genetic strategies hassulted in major advances in developmental biology, in pticular, our understanding of how specific genes influencepatterned movement of tissue.52 Nevertheless, these extraodinarily powerful strategies fall short when probes for cellar function need to be applied with high spatial~diffractionlimited! and temporal (!s) resolution. Laser microbeams53
have been used to surgically investigate the forces respsible for morphogenesis54,55 as well as to locally activategene expression.56 An infrared, near-field microscope thacombines spectroscopic and imaging capabilities has bused to investigate biological tissue as well as subcellustructures at the superconducting accelerator~SCA! FEL fa-cility at Stanford.57 Here we describe methods that optimibeam delivery to thick biological specimens while monitoing tissue response at high spatial and temporal resolutio
3224 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
FIG. 22. Optical layout of the laser microbeam facility. OF—optical flat, M—mirror, L—lens, Ir—iris, QTZ—quartz flat.
es.rodi
thalise
er
,
-
i-chthed onrly
re-ues
n,n itfol-the
inndre-
We have developed a laser-based microbeam to invgate tissue dynamicsin vivo at the Duke FEL LaboratoryThe system has been planned to accommodate numelight sources, including the UV OK-4 FEL and the infrareMark-III FEL as well as commercial laser systems. The mcrobeam can be used to perturb the tissue, cells, andconstituent biological molecules while monitoring in retime with confocal microscopy. High-resolution imagingachieved with laser confocal microscopy to visualize grefluorescent protein~GFP! labeling of the cytoskeleton intransgenic animals.
A. Microbeam
The microbeam is shown schematically in Fig. 22, whthe perturbing beam can be the IR or UV FEL, aQ-switchedNd:YAG laser system~Continuum YG571, 10 ns pulsewidth10 Hz! or a mode-locked Ti:sapphire laser system~Spectra-Physics Tsunamai, 100 fs pulse width at 80 MHz!. In thenear future the Nd:YAG will be replaced with the more com
ti-
us
-eir
n
e
pact Continuum MiniLite II. The layout of the confocal mcroscope is shown in Fig. 22 as well as in Fig. 23, whiconsiders the optical path of the perturbing beam withinmicroscope. Much of the early research has concentratemaking precise, user-defined incisions of fruit flies in eastages of development.55 For this, the commercial Nd:YAGlaser system allows progress in biological/biophysicalsearch while at the same time highlighting technical issthat can be generalized to FEL microbeams.
To minimize the spot size of the biological specimeeach perturbing beam must be of high optical quality wheenters the rear aperture of the imaging objective. In thelowing, we describe the optical system used with eithersecond ~532 nm! or third ~355 nm! harmonics of theNd:YAG laser system. The first stage of the optical trafilters unwanted wavelengths from the ablating beam aprovides fixed attenuation of the power. The beam firstflects from a fused silica optical flat~Janos Technology
FIG. 23. Schematic diagram of the microscope’s interior showing how the internal excitation beam and the external ablating beam are combindichroic mirror to simultaneously direct both onto the sample. Note that the ablating beam diverges slightly to compensate for longitudinal cicaberration of the microscope objective in the UV. A second dichroic mirror is used to separate sample fluorescence from light scattered by the excon andablating beams. The fluorescence is then focused through a pinhole~conjugate to the objective’s focal plane! before impinging upon a photomultiplier fodetection.
nd
c
r,
icr
iomd
32p
tef
mun
pin
oons
nuthoee
-lar-theofthe,
ser-
rolam
re-pa-thatave-ope
of. Tour-isofdi-
serstheti-
ab-G
ter
tingeen
A1805-358! 45° in the plane of the optical table. The secoharmonic iss polarized and the third harmonic isp polarizedwith respect to this optical flat. Thus the second harmoniattenuated to;1/20 and the third harmonic to;1/100 of theincident power. The beam then passes through a cologlass band pass filter~for 532 nm the Newport FSR-UG11for 355 nm, Thermo Oriel No. 51970! that cuts out any re-sidual light from the fundamental and unwanted harmonof the laser. For operation at 532 nm, the beam is thenflected from a second optical flat, providing total attenuatto ;1/400 of the incident power. For operation at 355 nthe optical flat is replaced with a dichroic mirror optimizefor reflection at 355 nm~CVI PAUV-PM-1037-C! to elimi-nate further attenuation at this wavelength.
The remainder of the optical train is used for either 5or 355 nm light. The ablating beam is passed through a stial filter to clean up the mode content. The spatial filconsists of two convex lenses~each with a focal length o75.7 cm for 532 nm light, CVI PLCX-25.4-360.6-UV! and a100 mm pinhole. We chose these long focal length lensesminimize the power density at the focus and thus avoid daage to the pinhole. The pinhole is held in a positioning mo~Melles Griot 07HPI501! on a small optical rail to minimizethe time required to reoptimize transmission through the stial filter when changing wavelengths. The size of the phole was chosen so it would pass only the center fringethe incident beam at the focus of the first lens. The seclens is used to recollimate the beam. This setup provideGaussian beam profile at both 532 and 355 nm.
To this point, the optical train has provided gross atteation of power and cleaned up the mode structure ofbeam. The beam is then transmitted to the optical tablewhich the microscope rests for fine tuning. As discussabove, the output of the Nd:YAG laser is polarized. We uscalcite polarizer~Newport 10GL08! mounted on a rotation
is
ed
se-n,
a-r
to-t
a--fda
-enda
stage~Newport RSP-1T! to provide continuously varying attenuation of the ablating beam. Immediately after the poizer, a glass coverslip is used to redirect a fraction ofablating beam to an energy meter for online monitoringthe energy per pulse. For the final conditioning stage,ablating beam is then passed through a pair of lensesf 1
520.0 cm~Newport SPX028! and f 2527.0 cm~CVI PLCX-25.4-128.8-UV!, on an optical rail. This pair of lenses idesigned to expand the ablating beam to fill the back apture of the microscope objective and to provide fine contof the divergence properties of the beam. The ablating beis now ready to be introduced into the microscope with pcise user control over its four critical parameters: power, stial mode, beam diameter, and divergence. It turns outsome divergence is necessary to compensate for the wlength dependence of the refractive index of the microscobjective.
We bring the ablating beam in through the Keller portthe Zeiss axiovert microscope of the 410 confocal systemallow simultaneous imaging and laser ablation, the first sface mirror that is mounted in the fluorescent filter sliderreplaced with a short pass dichroic mirror without lossimage intensity. The ablating beam passes through thechroic mirror and becomes coaxial with the scanning labeam~argon, 488 nm! of the confocal system. Both beamare then transmitted through the objective and ontosample. Different sets of dichroic mirrors are used to opmize the confocal fluorescent imaging performance whilelating at either the second or third harmonic of the Nd:YAlaser.
Two elements of the optical train are under compucontrol: a fast shutter~UNIBLITZ US25S2ZM0! just prior tothe spatial filter and the actuators~Newport CMA-12PP! onthe mirror underneath the microscope that steers the ablabeam towards the objective. Custom programs have b
n2
edllxs
rventovncp
die
thcili
na
ohethzeng
thedneth
thto
icam
a.bt
thres
sthengnsTonfois
ialmi-adyytheevan-
edpo-ialves.thethealamthetory.can.
lci-tion
ol-a-
sall assultenro-
d-
areages
as it
3226 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
written as Java plugins toIMAGEJ @National Institute ofHealth ~NIH!# and executed on a PC. The shutter is cotrolled through an RS-232 connection to a UNIBLITZ D12driver. In a similar way, the mirror actuators are controllthrough a second RS-232 connection to a Newport contro~ESP300-11N11N!. The shutter is used to control sample eposure to a defined number of pulses from the ablating laThe mirror actuators are used to scan the ablating beamthe two dimensions of the plane of focus.
We use confocal fluorescence microscopy to obsefruit fly embryos during laser microsurgery. The argon laswhich is internal to the microscope housing, is focused othe sample, exciting a cone of GFP molecules both aband below the focal plane of the objective. Fluorescefrom these molecules is collected by the objective and serated from scattered excitation light with a long-passchroic mirror. The fluorescence collected is then focusthrough a pinhole that is conjugate to the focal point ofobjective. This configuration largely rejects fluorescenfrom regions of the sample outside the focal plane. To buan image, the detector signal from the photomultipliercompiled as the excitation source is raster scanned.58
At the beginning of each set of experiments, the aligment of the ablating beam is carefully checked withrhodamine B dye in an agarose gel. The confocal microscis first adjusted to image the bottom surface of the gel. Twith the ablating beam blocked from the entrance tomicroscope, the spatial filter is optimized and the polariset to pass;200 nJ per pulse to the sample. While imagithe surface of the gel, the shutter is transiently openedallow single shots from the ablating beam to reachsample. The lesions produced in the agarose gel are usposition the ablating beam in the center of the image plaThe divergence of the ablating beam at the rear aperturthe microscope objective is then finely adjusted throughseparation of the final two lenses in order to minimizelesion size observed. As shown in Fig. 24, our abilitymeasure the absolute size of the lesion in soft biologtissues is limited because such tissues retract from thegins of the ablated spot. However, the lesions may besmall as 0.5mm in diameter when a metal film is targeted
Once the ablating beam has been optimized, we cangin a set of experiments. Typically one fly embryo is usedfinely adjust the energy per pulse in the ablating beam soit is just above the single-pulse ablation threshold. Fotypical experiment, we then take a single confocal fluorcent image of a fly embryo and open this image inIMAGEJ.59
The user then begins taking a time-lapse series of imagethe embryo and selects the desired cutting trajectory onembryo inIMAGEJ. On command, the system will then movthe two mirror actuators in order to position the ablatibeam focus in the image plane at the start of the user-deficut, open the shutter, move the ablating beam along the udefined trajectory, and close the shutter when complete.tasks of imaging and directing the ablating beam are ctrolled by two separate PCs. An example of real-time concal imaging during laser microsurgery of a fruit fly embryoshown in Fig. 24~C!.
-
er-er.in
er,oeea--deeds
-
pener
toeto
e.ofe
e
lar-s
e-oata-
ofe
eder-he--
B. Future prospects
We will need to have similar control of the power, spatmode, beam diameter, and divergence when adapting thecrobeam system to either FEL source. Since we are alreusing UV optics for the third harmonic of the Nd:YAG, onlsmall adjustments will be needed to use the OK-4 FEL asablating beam. Many more significant modifications will bnecessary to adapt the microbeam apparatus to take adtage of the preferential ablative properties of infrarradiation.8 We can simply replace the lenses and calcitelarizer with similar IR-transparent optics. However, speccare must be taken in the selection of microscope objectiWe could compensate for the chromatic aberrations inobjectives over the small wavelength ranges involved intripled Nd:YAG microbeam apparatus. To match the focplanes for a visible imaging system and an IR ablating bewe will need to use a reflective Cassegrain objective onmicroscope, which has already been tested in our laboraWith these modifications, the laser microbeam apparatusbe extended to take advantage of available FEL sources
IX. MARK-III FEL APPLICATIONSIN MASS SPECTROMETRY
Mass spectrometry~MS! is an indispensable analyticatool in biological, environmental, medical and polymer sences and technology because of its high mass resolu(;1024), wide mass range~up to 106 Da), and efficiency.60
A major challenge in MS is developing ‘‘soft’’ ionizationtechniques that leave large, thermally labile analyte mecules intact. Two competing complementary ‘‘soft ioniz
FIG. 24. Examples of laser microsurgery using the third harmonic~355 nm!of a Q-switched Nd:YAG laser.~A! By selecting a per pulse energy that liejust above threshold, the ablating microbeam can create lesions as sm0.5mm in diameter on solid samples like a metal film. Larger energies rein larger lesions.~B! A single pulse is delivered to the boundary betwetissues in a living fruit fly embryo. The three panels are a time-lapse pgression of confocal fluorescent images from before the laser pulse,,2 safter the pulse, and 14 s after the pulse. Within 2 s the lesion has expandeto a diameter of;2 mm. It continues to expand for tens of seconds following the laser pulse.~C! The ablating beam is steered so it cuts a lineincision (;75mm in length! across the entire width of an embryonic tissuknown as the amnioserosa. The panels are a time-lapse series of imcollected during this incision. The amnioserosa tissue retracts stronglyis cut.
tion’’ techniques are widely used in MS, electrospray ioniztion ~ESI!61 and matrix-assisted laser desorption ionizat~MALDI !.62 In MALDI, the analyte is encapsulated at loconcentration in a matrix that absorbs laser light. As the mtrix is ablated, the analyte molecules are entrained inplume of ablated material, cooled by free-jet-like expansand also, to some extent, shielded from collisions until ioization occurs.
The Mark-III FEL—with its combination of high intensity, high pulse repetition frequency, tunability, and ultrashpulse duration—offers two significant opportunities for eploring and improving the soft ionization process. First, aspectroscopic effect, such as ion yieldY, is proportional tothe energyE deposited per unit volumeV:
Y}~E/V!>F lasera~v,I !matrix, ~4!
whereF is the laser fluence,a is the absorption coefficientandv and I are, respectively, the laser frequency and intsity. Since many matrix materials have rich vibrational sptra in the midinfrared, the local density of vibrational exctation and the ability to initiate specific processes, suchionization, can be varied according to the wavelength aintensity. Second, the ionizationrate scales with the inten-sity, rather than the fluence
dN1
dt5h•N0s (k)~ I /\v!k, ~5!
whereN1 is the number of ions,h is the quantum efficiency~which can be taken to include all ion loss processes, succollisional neutralization!, s (k) is thekth-order cross sectionfor ionization, andI is the laser intensity. Thus in principlethe ionization rates associated with ns-laser-induced destion and ionizationshould belower than those for ps or fspulses; whether or not this is true in a practical laser ionition source for MALDI MS remains to be demonstrated coclusively; early FEL MALDI-MS studies showed clear evdence of intensity, rather than fluence, dependence;63 butmore recent experiments reveal a more complex picture64
A. Reflectron time-of-flight mass spectrometrywith the FEL
The Vanderbilt University FEL has been used to demstrate soft ablation and ionization of proteins with massesto 66 kDa from a variety of matrix crystals,65 so work hasnow been expanded to include such biologically relevant mtrices as water and gels,without adding any exogenous matrix materials. The Vanderbilt FEL has 1 ps micropulsspaced 350 ps apart in a macropulse lasting up to 4ms, withan average power of up to 3 W.6 Individual macropulsesfrom the free-electron laser are routed through a broad-bPockel’s cell12 to slice out a shorter segment that ranges fr100 to 400 ns, typically containing 50–300mJ in a few hun-dred to a thousand micropulses; the pulse energy is adjuby crossed polarizers. The laser spot size at the surface oMALDI target, measured with a reticle under an optical mcroscope, is typically 231024 cm2, producing micropulseintensities of 109– 63109 W/cm2. Laser energies are measured directly in front of the final focusing lens (BaF2 ,EFL510.6 cm) using a Molectron J25-110 energy meter a
-
-en-
t
y
--
sd
as
rp-
--
-p
-
s
nd
edthe
d
an EPM1000 controller. The readings are then correctedthe measured transmission of the lens and ZnSe windowthe vacuum chamber to arrive at the total energy depositethe target.
Mass spectra are acquired with a custom-bulaboratory-modified 3 m reflectron time-of-flight massspectrometer66 ~Comstock, Oak Ridge, TN!, equipped withdelayed extraction, a dual-chevron microchannel pl~MCP! detector~Galileo Electro-Optics!, and either a 96-well movable microtiter plate stage or a cryogenic samstage maintained near liquid nitrogen temperatures. Incryostage, the probe is slightly off the ion-optical axis of tmass spectrometer, so that one can ablate multiple spotthe target by rotating the probe. The temperature of the sis monitored by a resistive thermal device attached tofront of the cryostage by thermally conducting epoxy. Voage of 5 kV is applied 300 ns after laser irradiation to accerate ions generated in the evolving ablation plume; focusis achieved by an Einzel lens near the front end of the flitube. At the end of the ToF tube, ions are accelerated to 4just in front of the MCP. Ion signals are detected by a 5MHz data acquisition card~Signatec DA 500A! and pro-cessed by custom software based onLABWINDOWS CVI. Theapparatus is shown in Fig. 25.
A recent experiment showed the utility of the FELdesorbing and ionizing small proteins directly from a poacrylamide gel of the type used in electrophoresis septions. Angiotensin II~1160 Da! and bovine insulin~5637 Da!were dissolved in de-ionized water~high pressure liquidchromatography grade! containing 0.1% trifluoroacetic acidat concentrations of 5 and 3 mM, respectively; equal pa~v:v! of the two analyte solutions were then combined. Amm diam disk of the gel was cut out using a small punch aplaced in the sample probe. Approximately 4ml of the com-bined analyte solution was pipetted onto the gel disk; a10 min, the gel section was rinsed with de-ionized water a
FIG. 25. Schematic of the reflectron time-of-flight mass spectrometerperiment located at the Vanderbilt W. M. Keck Free-Electron Laser CenThe ion source is located in a large cubical vacuum chamber; diffesample holders, equipped, for example, with cryocooling or movable mititer plates, are mounted by interchanging the rear access plate ovacuum chamber. The FEL beam transport system, including the eleoptic switch, as well as other lasers~Er:YAG, N2 , KrF/ArF, and Nd:YAG!can be used interchangeably to provide laser ablation and ionization osamples.
end
tivthin
aen
se
o
y
f
oteic
-ble
gh
oarch
ag-ctede
mler-
leed
nfaceely
in ance
ea-uch
op-
o-
.to
ald
heions
hout
3228 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
blotted. The gel disk, mounted in the probe tip, was thsubmerged into liquid nitrogen for approximately 30 s aplaced in the cryostage.
Figure 26 shows a typical mass spectrum and the relayields of the angiotensin as a function of wavelength near5.9 mm vibrational resonance that includes both the bendmode of water and the first overtone of 12mm libration. Thedependence of the ionization yield on the wavelength isunambiguous signature of a resonant absorption procMore interesting, however, is the fact that the 10-fold stroger absorption resonance at 2.94mm produces no ionswhatever.67 While this null result agrees with previouMALDI studies using the Er:YAG laser at the same wavlength, no satisfactory explanation has been forthcomingdate. This result is particularly vexing since irradiationwater ice using the FEL at 2.94mm doesproduce ion signalsfor small proteins.68
B. Single-ion measurements of ion velocities
While the efficacy of IR MALDI MS has been clearldemonstrated using Er:YAG lasers,69 the mechanism of ion-ization remains a puzzle.70 To elucidate the mechanisms oinfrared laser-induced desorption and ionization,71 we mea-sured the flight times of single ions produced by ablationa conventional MALDI sample preparation that incorporaangiotensin into microcrystallites of 2,5-dihydroxybenzoacid ~DHB!.
FIG. 26. ~A! Mass spectrum of angiotensin and bovine insulin obtained5.9 mm from a polyacrylamide gel.~B! Wavelength dependence of the yieof angiotensin from the same polyacrylamide gel.
n
eeg
nss.-
-tof
fd
Samples of 2,5-DHB both with and without dilute concentrations of angiotensin were mounted on a translatasample manipulator in a small turbopumped ultrahivacuum chamber~Kimball Ion Physics!; base vacuum was1027 Torr. The FEL was tuned to 2.94mm, and 100 ns FELmacropulses were allowed to strike the MALDI target. Twexperiments were then performed. First, a Stanford ReseSystems RGA300 residual gas analyzer~RGA! was operatedas a quadrupole mass spectrometer~QMS! to identify unam-biguously the mass-to-charge ratio of the ions and ion frments from the laser ablation. Second, nonmass-seletime-of-flight experiments were carried out by allowing thcharged particles to drift in a field free region for 41 cbefore passing through a grounded grid and being acceated towards the channeltron~Burle 4730G! detector surface.The negligible effect of small axial fields in the quadrupomass spectrometer on the velocity of the ions was confirmby these time-of-flight~TOF! measurements. Single iopulses generated when an ion strikes the detector surwere measured in pulse-counting mode. This is an extremsensitive technique, since it avoids the dead-time effectstypical MCP; however, the small apertures at the entraand exit of the QMS, shown in Fig. 27~a!, keep a large per-centage of the ions from reaching the detector. Thus, msurements made using only the shielded detector are mmore sensitive because of a higher signal to noise ratio.
Detector signals are routed through a CLC401AJPerational amplifier~National Semiconductor! into the Sig-natec PDA500 500 MHz wave form digitizer, and then prcessed by customLABVIEW software, which controls theRGA, the PDA500, and anX–Y sample translation stageThe arrival time of each ion was recorded and was usedgenerate a TOF distribution (dN/dt), from which the energy(dN/dE) and velocity (dN/dv) distributions are calculatedby
t
FIG. 27. ~A! Experimental geometry for mass selected ion detection. Tsetup was used to determine the mass and energy distribution of cationemitted by FEL irradiation of DHB microcrystallites at 2.94mm. ~B! Ex-perimental geometry for higher sensitivity TOF measurements made witmass selection.
wherem is the mass,d is the distance from the target to thdetector,t is the detection time, andE5mv2/2. Appropriatedistribution functions are then used to fit the experimendata.
Figure 28 shows several features observed in the veity distributions produced in this extremely sensitive TOmeasurement.71 First, the velocity distributions of the DHBmatrix ions clearly reveal a fast and a slow component, bof which are too energetic to be explained by a thermmechanism. The fact that the velocity distribution is ainsensitive to the duration of the pulse, as shown bydotted line that indicates the DHB ion velocities for the fu4 ms macropulse, also indicates a nonthermal proceswork.
C. Future prospects
While the small, inexpensive N2 laser~337 nm! is ubiq-uitous in commercial MALDI spectrometers, the possibilof controlling the soft ionization process by appropriachoice of the laser wavelength and pulse duration canprinciple open up many new applications. For example,our laboratory we have shown that mass spectrometryorganic compounds in waste-storage tanks can be accplished by tuning the FEL to an absorption resonance ononorganic crystal, NaNO3, already present in the mixturethus avoiding the necessity of adding an exogenous mat
In particular, the increasing interest in imaging maspectrometry with high spatial resolution—on biochips,gels and other separation technologies, and in biologtissues—is likely to force reappraisal of the lasers now ufor MALDI technology. In clinical applications, for examplea chemical map of tissue or cells that provides the locationspecific biomolecules is clearly more useful than the mspectrum at an isolated point. Imaging techniques, howerequire high pulse repetition frequencies to be useful, wit
FIG. 28. Axial kinetic energy distributions of DHB and angiotensin derivfrom the TOF signals in Fig. 27. DHB has a relatively narrow distributiwhereas the angiotensin distribution is relatively constant from 0 to 3where it gradually decays to 0 at 6 eV. The broken line represents the endistribution of mass selected DHB1 generated by the full FEL macropuls~4 ms, 1.6 J/cm2).
l
c-
thl
e
at
innofm-a
x.s
ald
fsr,
1
kHz being a convenient benchmark. Therefore, it is likethat future analytical mass spectrometers will incorporsolid-state, high repetition-rate lasers, in spite of the adcost. Higher quality data acquired at higher speeds impcorrespondingly higher throughput and more sophisticadata interpretation, thus justifying the cost differential.
While free-electron lasers are too complex and costlyuse in routine analytical applications, research on soft iization processes using FELs can help to develop the spfications for solid-state coherent sources that produceable, ultrashort-pulse, high repetition-rate output neededfuture applications in mass spectrometry.
X. PUMP–PROBE TECHNIQUESFOR PROTEIN DYNAMICS
Here in Sec. X we review techniques to carry out timresolved pump–probe experiments in the infrared regionprotein absorption. The infrared is viewed in the broad senfrom 3 to 100mm. Time resolved is meant to apply to thtime domain where direct vibrational relaxation rates canobserved after excitation up an anharmonic vibrational lder due to an intense burst of ‘‘pump’’ IR photons. This timrange is typically 0.1–100 ps, times far too fast for mainfrared detectors. Since these times are so fast, the clasway by which to obtain time resolution is via pump–protechniques, which use optical delay lines to stagger the pupulse relative to the probe pulse.
There are substantial concerns regarding pump–prexperiments with proteins, however, due to the fact that~1!the sample is typically dissolved in water, a highly absotive liquid in the infrared, and~2! proteins have very smalnonlinearity in the infrared region and thus need very laintensities to achieve measurable pump–probe signals.72
We will not discuss the biological physics that canlearned from pump–probe experiments using the far-infraFEL at UCSB, the midinfrared SCA FEL at Stanford, alinac-based FELs,3,72–74but, rather, will discuss some expermental methods that have helped to solve the problems fawhen carrying out pump–probe experiments using FELsparticular we will discuss how one can use a time-delayprobe pulse~the ‘‘reference pulse’’! to decrease the commomode power fluctuations that are endemic to FELs andproblems associated with sample heating during a FEL mropulse.
A. Differential measurements
A typical FEL is not an externally seeded or regenetively amplified laser, but instead relies on amplified powfluctuations in the electron beam to seed the subsequenoutput. Not unexpectedly, the output of a FEL is subjectlarge energy fluctuations both within the profile of a singmacropulse and from macropulse to macropulse. As a rorule of thumb, one can expect the fluctuations to be at besthe order of 10% of the mean. The maximum transmisschange one can expect from a protein is at best 1% ipump–probe experiment, so extensive signal averagclearly is necessary. However, for copper-based linac FELis difficult to get repetition rates higher than tens of Hertz
,rgy
ch
foathi
th
cehep
isoiv
l
-e as
ng,:Ne
ed,ulseIn
ledaniast is
z.ac-ct-
lec-t itst
m-ledve-ry
os-
ye
ise.
the
3230 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
without some sort of common mode noise rejection tenique deep signal averaging is simply not time efficient.
Figure 29 shows a schematic of a free electron laserinfrared experiments macropulse. The amplitude of the mropulse can vary and there can be jitter in the start ofmacropulse from variances in the self-seeding of the IRthe optical cavity. A way has been devised to get aroundproblem, and the core of the idea is to optically delayreference pulse so that it falls in between the~nominallycoincident in time! pump and probe pulses. This referenpulse ~in the case of FELIX it is delayed 20 ns since tmicropulses come every 40 ns! has no corresponding pumpulse, but is identical to the micropulse that precededFigure 30 shows a possible configuration that producedelayed optical pulse. There are two key aspects of thistical layout:~1! the reference pulse is delayed 20 ns relat
FIG. 29. Basic shape of a macropulse from FELIX.
FIG. 30. Basic layout of a pump–probe experimental table with a delareference pulse.
-
rc-enisa
t.a
p-e
to the probe pulse, and~2! the spot size at the off-axis focaplane is 50mm full width at half maximum~FWHM!. Thiswas accomplished by the use of a 13telescope in the reference beam arm so that the reference spot size is the samthe probe spot size. All the optical elements are reflectiand no refractive optics were used anywhere so a Helaser accurately gave beam overlaps and focal points.
Once a comb of probe and reference pulses is obtainit is then necessary to invert and subtract the reference pfrom the probe pulse to eliminate common mode noise.the case of HgCdTe photoconductive liquid nitrogen coodetectors with bandwidths on the order of 20 MHz, this cbe accomplished by simply inverting the sign of the bcurrent for the reference pulse. Then the detector outpulow pass filtered, typically at a cutoff frequency of 1 MHFigure 31 shows a schematic diagram of how this wascomplished. Figure 32 shows the dramatic effect of subtraing the reference pulse from the probe pulse using this etronic scheme. The advantage of this technology is thaworks for any photoconductive detector of sufficiently fabandwidth, even in the far-infrared.73
There are, however, newer ways to achieve high comon mode rejection by using the new magnetically podetectors which offer subnanosecond resolution at walengths out to 12mm. These photovoltaic detectors are vefast and quiet, with detectivity (D* ) greater than108 cm (Hz)1/2/W ~Vigo Systems Ltd., http://www.psplc.com/!. Since these detectors are so fast, it is p
d
FIG. 31. Use of bias current oscillation to decrease common mode no
FIG. 32. Digitized traces of detector output with and without~labeled ‘‘Refblocked’’! the reference beam with bias properly phased to subtractreference pulse from the probe pulse.
sible to easily resolve individual micropulses within the maropulse train. Modern digital scopes such as the Tektro3052 offer 5 gigasamples/s, 500 MHz bandwidths, and Gtransfer rates of 100 000 bytes/s with a 9 bit analog to digitalconverter~ADC! and can be used to capture at 10 Hz putrains in the detectors. A ‘‘digital boxcar’’ has been deveoped at FELIX using the fast signal averaging and fast trafer rates of the 3052, the high speed response of the VIdetector, and a delayed reference pulse to do very fast boaveraging ofsingle micropulses, normalized to a referenpulse to compensate for fluctuations in intensity of the FEFigure 33 shows how the pulse train looks in a time-resolmanner for a single pump pulse, pump–probe experimusing this new technology. The reason for doing sinpump–pulse experiments is discussed in Sec. X B.
B. Energy and power considerations
One of the principal problems in performing biologicIR pump–probe experiments is the fact that the sample ually must be contained in water that strongly absorbs inand that large amounts of power are necessary to accoma significant nonlinear response in a protein. We should mtion here that the IR absorption of water vapor is alsomajor problem in IR pump–probe dynamics, but we hadiscussed this previously75 and assume that the experimenhas gone to great lengths to purge water vapor from all pof the IR transport system. In the case of liquid water,absorption length of water at 6mm is 4.6 mm,76 so it isnecessary to keep the sample thickness to 20mm or less.Further, in order to deliver the highest power for a givpulse energy it is important to keep the focused area ofpump pulse as small as possible. Fiftymm is an achievablespot size at 6mm with high f /number off-axis parabolic optics. This then means that the illuminated volume is qusmall, on the order of 531028 cm3. For a micropulse en-ergy of 10mJ, the rise in temperature is 200 K!
For 0.1 mJ per micropulse, we observe a more reasable temperature rise of 2 K per micropulse. However, aleast for copper linac FELs the IR pulses occur in a train
FIG. 33. Time-resolved pulse structure using magnetically poled detecThe probe and reference pulse before a single pump pulse are shownsample in this case was a silicon wafer, and the pump pulse was a 53pulse of visible light. The photogenerated carriers increase the reflectivithe silicon.
-ixB
e
s-Oar
.dnte
u-,
ishn-aertse
e
e
-
f
micropulses called a macropulse, and the rate at which thmicropulses occur is between 2 GHz and 25 MHz. Eventhe lowest repetition rate of 25 MHz that we have used this an accumulation of thermal energy in the focal region dto the finite rate at which heat can diffuse. There are mamisconceptions about thermal decay times. A common onthat heat is carried away at the speed of sound. This cabe true except in the case of shock waves, otherwise iftouch a hot object your feet would feel warm in a milliseond. In fact, heat dissipation is a diffusive process anddecay times are related to the diffusion times in liquidssolvent molecules. The equation for heat flow yields
¹2T5~Cp /k!]T/]t, ~7!
where Cp is the specific heat at constant pressure ofmaterial andk is the thermal conductivity. In our experiments, we have essentially one-dimensional heat flow, sthe sample thickness~12 mm! is small compared with the IRbeam diameter~50mm!. For analytical simplicity, we assumthat the initial jump in temperature caused by the laser iGaussian distribution like that expected for a TEM00 beam.Cooling of the sample is roughly exponential in time, and ttime constantt for cooling of the sample under these condtions is
t5L2Cp/4k, ~8!
whereL is the thickness of the sample. Table I gives coolitimes for various materials used in our sample constructassuming a 12mm sample thickness. We have added a cumn labeledtc , the time required for the sample to return5% of the original temperature. This time basically setsmaximum repetition rate for which no appreciable risetemperature will be observed in steady state.
Typically a macropulse from a copper linac FEL hasmacropulse duration of about 4 ms. Therefore, there is lited heat dissipation during a macropulse. The heat generfrom repetitive IR FEL micropulses is accumulated overmacropulse. With 1mJ energy per micropulse and 20%sample absorption, the jump in temperature from a sinmicropulse is 2 °C in H2O. With over 100 micropulses inone FELIX macropulse, this heat leads to an extremely larise in temperature of about 200 °C in H2O! This analysisshows that sample heating is a serious problem in pumprobe experiments. In order to measuretrue signals withoutcontamination due to sample heating, it is essential to ussmall number of micropulses at the beginning of each mropulse.
Another, more subtle, effect is due to the fact that inpump–probe experiment the pump and probe beams areherent with one another because they have been split f
rs.he
nmof
TABLE I. Time constants of thermal relaxation in selected materials.
3232 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
the initially coherent light beam coming from the FEL. Thmeans that the temperature pattern across the focal reginot uniform but instead is modulated in a sinusoidal mannThis thermal grating can diffract pump energy into the probeam in a constructive or a destructive manner, giving risoscillations in the apparent response of the sample topump energy which have very little to do with the actupicosecond relaxation of the vibrational levels.
It is easiest to see and document all of these purely tperature related phenomena by using a pair of boxcar igrators timed to look at the pump–probe signal at the bening of the macropulse when the sample is ‘‘cool’’ and at tend of the macropulse when the sample is thermodynacally ‘‘hot’’ due to deposition of energy into the samplFigure 34 is a schematic of the rise in temperature that caseen through a macropulse and the placement of boxcategrating gates that can be used to extract the cold andtemperature signals. Technically we should only use the leing signal from the macropulse signal, but it is interestingobserve the very strange effects that occur as the delaymoves the pump pulse relative to the probe pulse: the hgrating diffracted signal from the pump pulse is observedoscillate in phase relative to the probe pulse, resultingconstructive and destructive interference. Figure 35 shthe dramatic difference in signal seen at the front and reaa macropulse in pump–probe signals. Note that we dobelieve these signals to be true picosecond phenomenathey do not occur at the cold front of the macropulse, onlythe rear. Clearly, single pulse pump/probe experimentsmore preferable due to the sample heating problem.
We hope that in this review we have highlighted somethe techniques that make pump–probe experiments poson biological samples, and that discussion of the severe hing problem that must be understood and controlled in orto extract meaningful data was useful.
XI. APPLICATIONS IN TIME-RESOLVEDFLUORESCENCE AND TRANSIENT ABSORPTIONUSING THE SUPER-ACO FEL
The Super-ACO is a tunable, coherent source in thebased on a storage ring FEL.77 It produces 350 mW at 350
FIG. 34. Schematic of the rise in heat during a macropulse from FELThe lines are drawn with noise to try to indicate the typical variancespower that occur in a FEL macropulse. The boxcar gates show how puprobe signals can be extracted at the beginning and end of a macropu
isr.etoe
l
-e--
i-
bein-otd-onet-
ons
ofotncetre
fbleat-r
V
nm for the user community with 15–50 ps FWHM pulse arepetition rate of 8 MHz. The first applications experimenoccurred in 1993 with a study of the anisotropy decay ofcoenzyme NADH.78 Anisotropy decay results from perturbation of the initial distribution of moments of electronic transition due to Brownian reorientation and consequentlyscribes the rotational dynamics of the system. The goal iunderstand the thermodynamical equilibrium of differeconformational states of the molecule and their hydronamical volume in solution.
After the first one-color experiment using time-resolvfluorescence, a transient absorption experiment was deoped where the system is excited with the UV FEL andprobed by visible-UV absorption using synchrotroradiation.79 From microsecond flash photolysis to femtoseond laser photolysis, the principle of transient absorptspectroscopy is based on powerful optical excitation tpromotes a large fraction of the molecules into the excistate, which is then probed by a second white/tunable lisource. The transient absorption detection method is partlarly useful for the study of relaxation dynamics and direidentification of the associated excited states and/or transspecies via their spectral signature. A novel approach is tphoton spectroscopy based on the combined use of theage ring FEL and synchrotron radiation to study the eltronic states of various biological chromophores. Frombeginning of the development of microsecond flash photosis, biologists have investigated the chemistry of exogenor endogenous molecules of biological interest at differtime scales. These methods allow the investigation of funmental processes of chemical reactivity in the liquid phasnew wavelengths and over short time scales. For examcharge transfer~electrons, protons, radicals! or the relaxationof structure~isomerization, dissociation and recombinatiotorsion, movements of great amplitude, intramolecumovements! is involved in significant biological phenomensuch as photosynthesis, vision, structural modifications
.n–
e.FIG. 35. Difference in pump–probe signal between the front and rear ofmacropulse. Oscillations in the apparent pump–probe signal seenmovement of the delay line for the rear signal are presumably an artifacthe temperature-grating signal.
DNA or proteins, enzymatic reactions, mechanisms of traport, and photochemotherapy.
A. Time-resolved fluorescence experiments
The experimental layout of the time-resolved fluorecence experiments is shown in Fig. 36. The sample unstudy is irradiated by a short FEL pulse, and its fluorescedecay F(t) is recorded by measuring the arrival timesingle photons of fluorescence versus a synchronizationnal ~Fig. 37!. The photons must be selected randomly,more than one per excitation pulse, to avoid pile-up effeFor each fluorescence decay curve, about 10 million tcounts were stored, giving approximately 105 counts at thepeak ~Fig. 37!. The counting rate should be maintainedapproximately 10 kHz for optimum time response. Thusmeasurement is completed in approximately half an hoThe excitation wavelength was 350 nm (Dlexc51 Å beingthe linewidth of the FEL emission! and the emission wavelength was 460 nm (Dlem58 nm), selected by a H10 JobinYvon monochromator.
NADH ~b-nicotinamide adenine dinucleotide, reducform! from Sigma Chemical Co. was studied in a 10 mtris-~hydroxymethyl!-aminomethane buffer atpH 8, and con-tained 0.02% sodium azide as an antibacterial agent.final concentration of NADH in fluorescence measuremewas 18mmol/l, giving an optical density of about 0.1 at 35nm in the 1 cm path of a quartz cuvette. NADH can binda large variety of enzymes, and provide the H1 ions or elec-trons for chemical reactions, which are catalyzed by thehydrogenase. It constitutes anin vitro natural probe of theactive sites of these enzymes. In addition, it is used asindicator of the metabolic state of organisms and tissueimaging techniques. In its reduced form, NADH showsabsorption band at 340 nm and an emission band in theible, centered around 460 nm, with a quantum efficiency
FIG. 36. Schematic diagram of the photon counting fluorescence exment. A Hamamatsu R1564U-06 microchannel plate photomultiplierused for fluorescence detection, and a fast Hamamatsu S4753 silicontodiode with homemade amplification was used for synchronization.polarization of the incident light is rotated to the vertical direction byFresnel rhomb for detection at 90° in the horizontal plane of indepencomponentsI i andI' of the fluorescence. The polarized componentsI vv(t)and I vh(t) are obtained by orienting the emission polarizer to vertical ahorizontal positions, respectively. The apparatus functiong(t) was recordedat the excitation wavelength with a scattering solution of Ludox®~DuPontCo.! in place of the sample.
s-
-ere
ig-ts.al
tar.
hets
e-
nin
is-f
2%. The oxided form (NAD1) neither absorbs nor emits ithis spectral range.
The temperature dependence of the measured lifetdistributions of NADH fluorescence decay is shown in TabII. The decay is described well in all cases by three to foclearly separated relaxation processes, only two of whmake a significant contribution to the kinetics~0.28 and 0.62ns!. The temperature dependence of the nonradiative radescribed by an Arrhenius law, the frequency factorA andthe activation energyEa characterizing the dynamic quenching process. A linear fit, ln(1/t f2kR)5 ln A2Ea /RTassuming79 kR553107 s21 leads to an activation energy o2.660.3 kcal/mol and a frequency factor of 26131011 s21. We observe a shift of the lifetime amplitudefrom the long to the short component when the temperais increased, and Arrhenius dependence of both componwith similar activation energies of about 1.5 kcal/mol. Tamplitudesci of the different components in Table II can, tfirst approximation, be identified as the relative populatioof chromophores that have the corresponding fluorescelifetime, with the equilibrium constant beingK5c2 /c1 . Thethermodynamic parameters~Table II! governing this equilib-rium are obtained by the linear fit
ln K52DH
R S 1
TD1DS
R, ~9!
assuming that these parameters are approximately constathe temperature range studied. These data are in good ament with data obtained by other techniques.80
Polarization of the fluorescence depends on the distrtion of moments of the electronic transitions responsible
ri-sho-e
nt
d
FIG. 37. Single photon counting data obtained with the Super-ACOFEL: ~A! Total fluorescence decayF(t) of NADH at pH 8, 10 °C,lexc 350nm, lem 460 nm, and corresponding instrumental functiong(t); ~B! residu-als obtained from maximum entropy method analysis of the decay in~A!;~C! autocorrelation of the residuals.
3234 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
TABLE II. Fluorescence measurements of NADH performed with the super-ACO FEL.
the emissions process, which is related to rotational Broian motion of the molecules in solution. The decay of anisropy of a chromophore rigidly attached to a sphere candescribed byr (t)5r 0 exp(2t/t) wherer 0 is the static anisot-ropy related to the angles between the moments associatabsorption and to emission andt is the relaxation coefficientwhich is inversely proportional to the rotational diffusiocoefficient of a sphere. The measurements of fluoresceanisotropy decay lead to very fast depolarization of the nitinamide ring, independent of the rest of the NADH moecule. We find an average hydrodynamic volume of 106100 Å3 for NADH, which would correspond to a spherwith radius of 6.2 Å. This is in good agreement with thvolume of a folded configuration (836 Å3) given by a vander Waals model. The spatial conformations deducedshown in Fig. 38 in folded and open conformations. Theresults are in agreement with those obtained from previnuclear magnetic resonance~NMR! studies of aqueousNAD.81
B. Transient absorption experiments
Figure 39 shows the transient absorption experimesetup employed at Super-ACO. Control of the spatial overbetween the pulses is critical. To monitor this, we mountequartz lens, which was used to focus the pump beam onto
FIG. 38. Possible conformations of aqueous NADH.~A! Folded form con-structed from minimized fragments of the NADH chemical structure unstereochemical rules borrowed from the DNA structure.~B! Structureadopted by NADH within the active site of lactate dehydrogenase internary complex with oxamate, according to the x-ray crystallographic dof C. Abad-Zapatero, J. P. Griffith, J. L. Sussman, and M. G. Rossm~1987! J. Mol. Biol. 198 445, entry 1LDM of the Protein Data Bank.
-t-e
to
ce-
0
rees
alpahe
sample, on anX–Y–Z manipulator. The pump beam cathen be moved and focused precisely in the region whereprobe beam intersects the sample. This region where thephoton beams intersect is optimized with the signal obtaiby a charge coupled device~CCD! camera in steady-statoperation. The time-resolved experiment was performedlower repetition rate, i.e., 83.2 kHz, than the normal o~8.32 MHz! using a Pockel’s cell.
Acridine from Prolabo, used without further purificatiowas studied. Absorption spectra were recorded using a C210 ~Varian Inc.! spectrophotometer. Typically, a 0.1 cm or1 cm quartz optical cell was used. Its final concentration w116 mM and 1.6 mM, giving an absorbance of about 0.5350 nm and absorbance of about 1 at 363 nm, respectivelthe 0.1 cm path of a quartz cuvette.
The absolute intensity of a differential transient absotion spectrumDA(l,t) is directly proportional to the population of the molecule in an excited state. Assuming thatpump beam interacts only via the transitionS0→S1 , thesaturation fluence isFS51/sa0(l) wheresa0(l) is the ab-sorption cross section of the ground state~in units of cm2) atwavelengthl. The pulse fluence of the pump should bethe same order as the saturation fluence in order to exclarge fraction of the molecules. The FEL power on tsample is limited to a maximum of 100 mW, which corrsponds to energy of 12 nJ per pulse~i.e., 231010 photons/pulse!. The pump beam is focused onto the sample to wita diameter of;20mm, leading to the possibility of fluenceapproximately equal to the saturation fluence. In the simpsituation, the pump/probe signal observed is proportionathe population that is not in the ground state. Thus, ifground state is the only absorbing state, then the pump/pobservable is given byS(t)5A^Pex(t)& whereA is a con-stant determined by experimental parameters such as thser intensity and the probability that the system is foundthe excited statePex&. If the probe wavelength is in thespectral region of existing absorption of the ground stateof the stimulated emission, bleaching of the solution mayobserved. This was the case in our investigation of acridThe change in transient absorption shown in Fig. 40 displdecay of the transient at 430 nm. The lifetime of the fi
singlet excited state determined by convolution is 1(60.2) ns. These results are consistent with those of Shaand Winn82 ~fluorescence lifetimes of 375650 and 817680 ps for excitations at 355 and 396 nm, respectively!.
C. Future prospects
The energy per pulse of an advanced storage ring Fsuch as the one proposed for the French Synchrotron Ration facility SOLEIL will be sufficient~3–10mJ! to carry outquasisaturation of the excited state in the active volum
FIG. 39. Schematic diagram of the transient absorption experiment~SU7—undulator; SB5—bending magnet; PD—photodiode; PM—photomultiplBS—bunch signal!. The UV-SR FEL pump was variably delayed bycomputer-controlled translation stage, with accessible delay ranges14.5 to 23.5 ns ~a positive delay corresponds to the UV-SR FEL pumpulse impinging on the sample before the probe pulse!. The white lightemitted by a bending magnet via the SB5 beamline is extracted to gena white probe pulse in the 400–700 nm range. Both the pump and pbeams are focused onto the sample, and cross at a 7° angle in the sawhich flows through a 1 mmthick quartz cell. The entire white continuumpulse is sent to an imaging spectrometer~Princeton Instruments, Inc. modeSpectraPro-750 monochromator! and detected by a thermoelectricalcooled CCD camera~TE/CCD-1752-PF/UV, chip size: 17523532) insteady-state operation or by a photomultiplier~Hamamatsu RS928! in time-resolved operation.
1iro
Lia-
e.
Moreover, this capability will allow operation to the edgevacuum UV~200 nm or possibly shorter!, which will openan entirely new spectroscopic field for transverse acou~TA! spectroscopy in the nanosecond and subnanosecongimes. Thanks to the tunability of advanced storage rFELs, it will be possible to study the excited states producby UV irradiation of a wide variety of chromophores, indcated in Table III. Such photoreactions define the very fistages of certain processes of cytotoxicity.84 Besides, the pri-mary species that result from the excitation of trytophanmainly the origin of the photodegradations induced byirradiation of proteins. In addition, nonradiative relaxationstill poorly understood, i.e., specifically how it affects thintensity and the kinetics of tryptophan fluorescence, whis used extensively in the study of structural and dynamproperties of proteinsin vitro.
The exact photophysical consequences of the excitaof the peptide bond (lexc5200 nm), present at very stronconcentrations in proteins, are still poorly known. Themechanisms could be of great importance for the comphension of complex evolutionary processes such as the aof proteins. Oxidative condensation of catecholamines s
;
m
atebeple,
FIG. 40. Pump/probe data for the acridine in ethanol. The smooth lineresents the fit of the data.
TABLE III. Intrinsic and extrinsic chromophores in biological systems. Foa list of additional extrinsic chromophores~used in fluorescence! see, forexample, Ref. 83.
Coenzymes and prosthetic systemsNADH, NADPH ~350 nm!, flavins ~450 nm!
Hemes: 400 nm, 550 nm and porphyrins: 500–650 nmQuinones: 260–280 nm, retinal: 280 and 500 nm
Pigments and neurotransmittersChlorophylles~360, 580, and 800 nm!, carotenoids~400–500 nm!
Adrenalin ~280 nm!, adrenochrome~300 and 480 nm!
Nucleic acidPurins and pyrimidins~260 nm!, wybutine~310 nm!
DrugsRadiosensitizer, psoralens, phenothiazins
feo
foTansena
toocar-
o
Eath
soheonmis
ic
mig
i
ieohrrtiic
oa
umthn
ro
a
nem
90°
ineionsta-
m
tedheff-sc-nd
s ofir-
irroril-IR.lrror
ro-thehisnceal
r-
-oning
ee-is
edthatlseineter.
nd
on-he
thehiss ass atared
2-4
3236 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
as adrenalin (lexc5280 nm), implied by the formation ovarious forms of melanin, brings into play uncharacterizreactive species, which transient absorption spectroscpromises to identify.
The nanosecond time scale constitutes a ‘‘key’’ fieldcomprehension of the primary photophysical processes.ing into account the intensity of the electronic transitioS0^2&S1 , the intrinsic lifetime of the excited state for theschromophores is in general about a few tens of nanosecoThe various decay processes of this excited state, whichthe starting points of the various photochemical or phophysical pathways, must thus be on an equal scale or mrapid than the nanosecond. Thus accessing subnanoseprocesses holds the promise of reaching the initial crossrofrom which the evolution of the excited state will be detemined.
XII. UV-PUMP, BROADBAND-IR-PROBESPECTROSCOPY
The Duke storage ring, OK-4 FEL, is a pulsed sourcecoherent UV radiation, tunable from 193 to 400 nm.85 Inaddition, the bending magnet downstream of the OK-4 Fis a source of broadband infrared radiation. Because the sbunch of electrons emits both pulses, the timing betweenUV and IR pulses is essentially jitter free and the time relution attainable is thus limited only by the pulse width. Tcapability to excite systems with tunable ultraviolet radiatiand then probe the relaxation processes throughout theinfrared with a time resolution on the order of 100 psunique, and it opens the way to the study of photochemand photobiological systems not previously accessibletime-resolved infrared spectroscopy. A beamline was comissioned to take advantage of these synchronized lsources, recognizing that the repetition rate of 2.79 MHzideal for rapid-scan, asynchronous sampling.86
A. Synchronized light sourcesfor two-color, time-resolved spectroscopy
The OK-4 FEL has been operational since 1996 and1999 water-cooled copper disks were installed in the cornof the storage ring with optical flats that extract synchrotrradiation. The acceptance angles for the capture of synctron radiation are 57 mrad horizontally and 14 mrad vecally. While the vertical acceptance angle severely restrthe collection of the far infrared (,500 cm21), theoreticalextraction of midinfrared radiation is comparable to thatbeamlines at National Synchrotron Light SourceBrookhaven National Laboratory.87
Figure 41 shows the layout for the optical and vacusystems to collect infrared synchrotron radiation, wherecopper flat reflects radiation downward. A gold-coated plamirror located 20 cm below the flat reflects the synchrotradiation towards a 6 in. diam,f /6 spherical mirror. Thespherical mirror directs the synchrotron radiation back pthis plane mirror and focuses the light through a CaF2 win-dow that separates the ultrahigh vacuum (10210 Torr) of thestorage ring from the remainder of the beamli(1027 Torr). The synchrotron radiation that emerges fro
dpy
rk-
ds.re-reondds
f
Lmee-
id
alto-
hts
nrsno--ts
ft
een
st
the storage ring vacuum chamber is then collimated by aoff-axis parabolic mirror (f eff57.5 in.). The collimated syn-chrotron radiation is then directed 20 m down the beamlby a pair of plane steering mirrors. The synchrotron radiatis brought out of the beamline vacuum system at the endtion through a second CaF2 window. At this point, the syn-chrotron radiation ~SR! encompasses wavelengths fro;450 nm to 9mm, limited at long wavelengths by the CaF2
windows.Once at the endstation, Fig. 42, the SR is direc
through a N2 purged path to the same level as that of toptical table by a steering mirror and focused by a 90° oaxis parabolic mirror (f eff55.5 in.). The focal point serves aan effective point source for a Bruker IFS-66v FTIR spetrometer. We have measured the SR power with FTIR acompared it to theoretical expectations. Due to the effectbeam divergence over the 20 m between the collimating mror and the endstation and reflection losses from each mand window, approximately 40% of the theoretically avaable synchrotron radiation power is delivered to the FTUsing a collimating mirror with a longer effective focalength could reduce divergence losses. However, the micurrently in use has the longestf eff available in a diamond-turned 90° off-axis paraboloid.
The storage ring current and consequently the synchtron radiation power decrease exponentially duringcourse of a single injection cycle of the storage ring. Tcreates a drift problem for measuring pump–probe differespectra, since the low light levels require 5 min of signaveraging to achieve a signal to noise ratio~SNR! of 1000~i.e., the ability to detect pump-on versus pump-off diffeence signals of 0.001 absorbance units!. Consequently, wehave installed an electronic shutter to measure the pumpand pump-off spectra as an interleaved set, thus avoiduntoward consequences of drift in optical power.
To perform the two-color pump–probe experiment wmust also deliver the coherent UV output of the OK-4 freelectron laser to the sample, as shown in Fig. 42. Thisaccomplished by extracting the UV light from its evacuatbeam tube that is mounted parallel to the vacuum tubetransmits the infrared radiation. Once the UV-pump puarrives at the optical table, it is directed through a delay land then focused onto the sample in the FTIR spectromeWe can vary the relative arrival times of the UV-pump aSR-probe pulses at the sample by.10 ns, at which the in-tensities of the two pulses have been shown to remain cstant within 5% without any realignment of the optics as toptical path length is varied.
A Hamamatsu streak camera was used to measurepulse width of the synchrotron radiation. The results of tmeasurement show that the SR pulse width increasesingle-bunch current in the ring is increased, up to 360 p10 mA. When these pulse width measurements are compto those taken in 1998 over a lower current range~the twosets of measurements overlap near 3 mA!, the pulse widthsare now a factor of 2 higher than they were in the past.88 Byoptimizing the electron orbit in the ring, this factor ofshould be recovered. The lasing pulse widths of the OK
FIG. 41. Schematic diagram of the broadband IR collection optics of the synchrotron beamline.
n
anslyeh
eleeth
is
plezedof
or aen-anrtur-hetheateareobesonen-ro-ter-
will be a factor of;5 smaller than the synchrotron radiatiopulse widths.
We have commissioned this pump–probe beamlineperformed initial characterization of both light sourceWhile the synchrotron radiation power level is on;200mW between 2 and 9mm, we have maintained thintrinsic brightness of the SR throughout the beamline. TSR pulse width is 200–400 ps~FWHM! depending upon theamount of current stored in the electron bunch. Coherradiation from the OK-4 that is used to excite the samphas a pulse width 2 – 53 shorter than the SR and approach5 mW of average power for single-bunch operation ofstorage ring.
B. Asynchronous sampling
The high intrinsic repetition rate of this storage ringcompatible with the asynchronous sampling method.89 For
d.
e
ntsse
our applications, the broadband IR source probes a sammaterial that is resonantly pumped in the UV, as summariin Fig. 43. The IR source can be viewed as a distributionFourier components. Part A represents an interferogram fsingle Fourier component, i.e., one first modulated in intsity by the Michelson interferometer, then passing throughabsorbing sample onto the detector. Part B represents pebation to the probe signal due to repetitively pulsing with tUV pump. In part C, a pulsed IR probe is synchronized toUV pump to interrogate the relaxation of the excited stdue to UV absorption, where the traces in parts A and Bincluded for reference. Since neither the pump nor the prpulses are synchronized to the scanning of the Michelinterferometer, over time all phases of the modulated intsity are measured. As shown in part D, the discrete interfegram of part C has been converted into a continuous in
onr
t.ouey
be,nchtheat
er-the
in
ine
heioneenis
o-roseshe, thegerhedis-a
ni-si-inen-heata-
s-n-entyn-the
ate,
-
willK-4
lyusd aof
enex-
ereet.rb-or
an
lote
3238 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
ferogram with a low pass electronic filter. Note the repetitirate must exceed the highest Fourier frequency by a facto2 ~40 kHz in our case! to avoid aliasing, a sampling artifacFourier transforming the sum of the interferograms for allthe Fourier components results in a single-beam spectrwhich can be put into a ratio against the no-excitation sptrum to generate a differential absorption spectrum. By s
FIG. 42. Schematic diagram of the endstation for UV-pump, broad bIR-probe spectroscopy.
FIG. 43. Asynchronous sampling method for time-resolved FTIR that pthe intensity as a function of mirror retardation or, equivalently, of time. Sthe text for further explanation.
of
fm,c-s-
tematically varying the delay between the pump and prothe overall relaxation processes are mapped. In single-bumode, the synchrotron radiation pulse repetition rate ofDuke storage ring is 2.7898 MHz, well in excess of threquired for asynchronous sampling FTIR.
C. Future prospects
As an example of the type of experiments to be pformed, consider the unanswered questions that surroundenzymatic mechanism of DNA photolyase. This protebinds to UV-induced lesions in DNA~specifically pyrimidinedimers! and then catalyzes the cleavage of the pyrimiddimers when exposed to blue light.90 All evidence points tothe involvement of photoinduced electron transfer in tcatalytic mechanism. However, time-resolved absorptspectroscopy in the UV–visible spectral region has not bable to identify the nature of the intermediates in thprocess.91 Once the photoexcitation of the bound flavin chrmophore was quenched, an unidentified intermediate awithin 2 ns of the flash (lmax5400 nm). The remaining stepwere silent in the visible spectral region. By extending taccessible spectral range to include the near and mid-IRelectron transfer events in this photocycle should no lonremain spectrally silent. Vibrational frequencies in tmid-IR are extremely sensitive to changes in the spatialtribution of electron density. By simultaneously probingbroad range of mid-IR frequencies, we will be able to motor the catalytic involvement of the flavin chromophore, redues of the enzyme, and the DNA bases of the pyrimiddimer. In addition, the time-resolved FTIR spectra will cotain information regarding conformational changes in tprotein and DNA backbones that accompany the photoclytic cycle.
It is instructive to estimate the minimum power necesary for the proposed investigation of DNA photolyase. Cosider an optimized time-resolved pump–probe experimwith parameters that match the Duke OK-4 FEL and schrotron beamline. Using the synchrotron radiation asprobe with a single electron bunch in the storage ring (f rep
52.79 MHz), our sensitivity is limited toDAIR,min50.001.For a strong vibrational band, e.g., an amide or carboxylwe have« IR5600 M21 cm21. If we optimize the optical sys-tem to provide a diffraction limited spot in the Bruker IFS66v sample compartment thenw550mm. The power re-quired decreases as the wavelength increases, so wecalculate the power necessary at the upper end of the Otuning range,l5400 nm. Furthermore, we will allow thesample to absorb all available light. Actual samples will onabsorb between 50% and 90% of incident light, and thincrease power needed. Under these conditions, we finthreshold power of 109 mW for an excitation wavelength400 nm~i.e., 40 nJ per pulse at 2.79 MHz!. Several hundredmilliwatts on the sample is a judicious target threshold givthis idealized assessment. One option is to consider anperimental system with more intense absorption, wh600 M21 cm21 is about as strong as vibrational bands gAlternatively, electronic transitions can exhibit molar absotivities 1003 greater. In fact, a HgCdTe semiconduct
FIG. 44. Block diagram of the monochromatic x-ray source built and currently operating at the W. M. Keck FEL Center at Vanderbilt University.electron gun; linac—linear accelerator; IZ—interaction zone; Dump—electron beam dump; the tabletop terawatt laser consists of a pump, yttriuhiumfluoride, stretch/regen, seed, Nd:glass amplifier, and pulse compressor.
th
ucgeurv-mtim
ndsste
IheIIara
-a
To
ailinmIRp
tho
emin
ur-ble,
il-
r isz
omsed
e
ap-bi-
rainres-n ofee-ercs.hehe
served as the experimental system for commissioning ofNSLS time-resolved IR beamline.87
XIII. IMAGING WITH PULSED, TUNABLE,MONOCHROMATIC X RAYS
Many individuals have long sought the means to prodpulsed, tunable, monochromatic x rays at high flux in aometry suitable for human imaging, as well as for other pposes. While ‘‘hard’’ x rays are currently available from seeral sources, they have never before been available frocompact source that allows one to control the spectrum,ing, and flux of the x rays simultaneously.
The phenomenon of inverse Compton scattering leitself well to the production of such a beam. In this procea high-energy electron beam is tightly focused and counpropagated against a powerful infrared laser beam. Thephotons scatter off the electrons and are shifted from theto x-ray frequencies in a direction almost collinear with tdirection of travel of the electron beam. Since the Mark-FEL produces a powerful IR beam and is driven by suchelectron beam, it has shown promise as a vehicle for x-production using the Compton process.
In August 1998, the Mark-III FEL at Vanderbilt University successfully produced pulsed, tunable, nemonochromatic x rays.92 That experiment yielded 104 x-rayphotons/s, deemed impractical for the uses envisioned.extreme radiation environment around the FEL, the lossflux through a mosaic crystal transport system to an upstshirtsleeves imaging lab, the inaccessibility of the beamcomponents in a thick concrete shielded vault, the low nuber of photons per pulse, and the downwardly spiralingoutput of the FEL as electron beam parameters were omized for the monochromatic beamline component ofx-ray source, all underscored the need for a better, mcompact dedicated system.
e
e--
a-
s,r-IRR
Iny
r-
hef
rse-
ti-ere
To that end, a new monochromatic x-ray imaging systwas designed and built. It consists of a rf linac running‘‘single pulse’’ mode, and a tabletop terawatt laser~Fig. 44!.This source became operational in April 2001, and is crently used for applications research in pulsed, tunamonochromatic x-ray imaging.
A. Monochromatic x-ray source
1. The accelerator
The electron gun is a copper photocathode, which isluminated by a portion of the seed laser output~describedbelow! that has been quadrupled. The linear acceleratomade of a single SLAC section driven by the 2856 MHoutput of a standard klystron. Its energy can be tuned fr20 to 50 MeV. One superconducting solenoid magnet is uto focus the electron beam to a 50mm spot at the interactionzone~IZ!, which is defined as the point of interaction of thelectron and IR beams.
2. The laser
The tabletop terawatt laser consists of a 200 fs Ti:sphire seed laser running at 1052 nm which drives a comnation stretcher/regenerative amplifier that produces a tof pulses stretched to about 1 ns, a Nd:YLF pulse compsor, and a frequency quadrupler that uses a small portiolight from the amplifier to drive the photocathode of thaccelerator, with the rest of the light from the amplifier dlivered to a multistage Nd:glass final amplifier to deliv20 J of IR light to a pulse compressor and focusing optiThese in turn deliver the final 10 J pulse to the IZ. Tcurrent repetition rate of the laser is 0.01 Hz, limiting tx-ray pulses to one ‘‘burst’’ each 100 s.
ngle0g
nIR
imt
akticinre
octhtuc
esonso
e
iathineeaonmh
tioa
vera
adin
aga
veinreo
ly
b
enas
pro-neer,
rasts.’’ishern
k-ng.er-dif-e upen,
0le-ard
s ofestsove-ticinto
ts atedgare
of-yarede-
redred. Byps,ordsin
der
beuchses
p-al
ed-rayms
3240 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
3. The integrated x-ray source
The integrated x-ray source counter propagates a si8 ps electron pulse containing 1 nC of charge and a singps 10 J pulse of 1052 nm IR light at the IZ to produce 110
x-ray photons, tunable from 12 to 50 keV with a varyinbandwidth of 1%–10%, also in 8 ps.
Alignment of the two beams is accomplished using cojoined perpendicular screens for serial visualization of thebeam and transition radiation from the electron beam. Ting and phase adjustments bring the Rayleigh ranges oftwo beams into an overlapping configuration at the IZ, ming use of a newly developed phase corrector and an optrombone. X-ray output emanates through a beryllium wdow on the end of the vacuum line. No x-ray optics aneeded for deflection of the beam unless one desires to for deflect the beam for certain experiments. By alteringenergy of the electron beam, the x-ray energy is madeable. Focusing the electron beam to a smaller or larger fospot size will vary the bandwidth. Since the machine issentially self-shielded, it does not require the use of a ccrete vault and may be run in an occupied room near pernel not required to wear radiation badges.
B. Medical applications
1. Mammography
Tunable, narrow bandwidth x rays can significantly rduce the radiation dose delivered to a patient in any typex-ray procedure currently performed. This savings in radtion dose can vary by a factor of 2–50, depending onstudy being performed and the imaging protocol beused.93 An example of the utility of such a beam would btunable, monochromatic x-ray mammography performwithout breast compression to yield three-dimensionvolumetric CT images. Such studies would unravel the cfusing overlap in structures now seen in plain film geoetries. Diagnostic accuracy of this type of mammograpshould theoretically rise due to the higher linear attenuacharacteristics of malignant tissues relative to normal brestructures.94 An x-ray source, such as that described abocan be configured to service a large multiroom mammogphy facility.
Images of breast phantoms are currently being mwith single 8 ps bursts of x rays, since each pulse contamore than enough photons to produce a complete imGenetically engineered mice are also being imaged ineffort to discern the development of tumors of the breast otime and the potential for reversal of these growths usnew types of drug therapy. The low radiation doses deliveusing monochromatic x-ray beams allow the performancelongitudinal biological studies without ‘‘frying’’ the mousewith the high radiation doses now needed that utilize pochromatic beams.
2. K-edge imaging
K-edge imaging becomes possible with such a tunabeam as well. By tuning to the binding energy of thek-shellelectron in a whole host of atoms, one can selectivelyhance the visibility of currently used radiographic contr
le8
-
-he-al-
usen-al--n-
-of-eg
dl/-
-ynst,-
ese.nr
gdf
-
le
-t
agents, as well as tumor seeking drugs labeled with appriate atoms within the 12–50 keV range of the beam. If owished to lower the radiation dose to the patient even furthgadolinium containing magnetic resonance imaging contagents could be substituted for iodine containing ‘‘dyeThe k-edge of iodine is 33 keV, while that of gadolinium50 keV. The human body is more transparent to the higenergy x-ray beam, but thek-edge effect can still be takeadvantage of.
3. Phase contrast imaging
The x-ray output of the source is easily collimated, maing it useful for the performance of phase contrast imagiThis type of imaging takes advantage of the inherent diffences in refractive index of various body tissues and offractive edges at tissue planes. Because the body is madof low atomic number elements, such as hydrogen, oxygcarbon, and nitrogen, phase contrast effects~which are bestseen in lowZ elements! can offer as much as 100–100times the information than would be derived from such ements solely using the linear attenuation effects of standabsorption imaging.95 This wealth of additional informationoffers some rather spectacular enhancements to imageanimals, tissues, and potentially, it is hoped, humans. Tusing cancers in whole excised breasts have shown imprments in the conspicuity of some of the stellate fibrostranding around small tumors, making them more visiblethe laboratory setting. In other words, the body reactsmany tumors by depositing fibrous tissue, which presencharacteristic starburst pattern in x-ray images. It is expecthat this novel capability will be added to the clinical settinonce some of the logistics of phase contrast imagingworked out.96
4. Time-of-flight imaging
Since imaging fluxes are produced in 10 ps, time-flight imaging could be performed with this source. X-raphotons that traverse the imaged part without scatteringtermed ballistic photons. These produce an image on atector within picoseconds of the start of imaging. Scattephotons will not reach the detector before several hundpicoseconds, and can extend into the nanosecond regimeusing a detector that only records x rays for about 100one can ignore the delayed scattered photons. This affone the opportunity of improving the signal-to-noise ratioan image by six- to ninefold.97 Alternatively, one could per-form an image with the same S/N ratio with almost one orof magnitude fewer X rays.
5. Exceedingly high speed imaging
While an exceedingly high speed pulsed beam wouldextremely useful in imaging rapid mechanical processes sas nondestructive testing of turbines, or explosive proceslike the study of modes of armor failure with kinetic weaons, it can still find medical uses in the areas of small animimaging or human imaging. Biological processes are exceingly slow compared to the picosecond structure of the x-pulse. However, splitting the beam into a number of bea
and directing them through an animal from different diretions at once opens the door to single pulse 10 ps CT. Twould allow the study of these creatures without the necsity of anesthesia. Imaging equipment could be simplifisince gating studies to cardiac and respiratory cycles wono longer be needed. Tunable x rays would also cut thediation dose needed to carry out these studies, so anicould be studied repeatedly over time to follow tumgrowth, drug action, or other disease processes.
C. Protein crystallography
This integrated x-ray source can also be used to perfprotein crystallography without the need to take protein crtals to synchrotron facilities. The beam can be focused toexceptionally small spot and, when run in a high averapower mode at 20 Hz, can deliver a photon flux only onetwo orders of magnitude lower than that typically deliverto the crystal by the synchrotron source. Higher fluxesavailable at synchrotrons but are frequently not used dudamage to the crystals. This device can deliver 8–50 kwith little modification at narrow bandwidth, making it possible to perform standard crystallography, multiwavelenanomalous dispersion~MAD !, and Laue studies using mutiple energies simultaneously.
D. Future prospects
While the current integrated x-ray source has been bto cover the x-ray spectrum from 12 to 50 keV, it is easscalable to higher energies. The addition of a second SLsection allows electron beam energies of 50–100 keV, anon.
XIV. SASE FELS AND POTENTIAL APPLICATIONTO BIOLOGY
A. Self-amplified spontaneous emission
When an electron bunch traverses an undulator~i.e., aperiodic magnetic field!, it emits electromagnetic radiatio~Fig. 45! at wavelengthl r5lu (11K2/2)/2g2, wherelu isthe undulator period,gmc2 the electron beam energy, andKthe dimensionless undulator strength parameter. Sinceelectrons are in a bunch much longer than the radia
FIG. 45. Schematic diagram of a single pass FEL operating in SASE mThe microbunching process that, develops in parallel with the radiapower is shown in the lower part.~Reproduced with permission from Re104.!
-iss-,lda-als
m-neo
etoV
h
ilt
Cso
hen
wavelengthl r , the EM field emitted by different electronhas a random relative phase. The incoherent EM wapropagate through the undulator and interact with the etrons. The interaction makes the trajectory of electrons wlarger ~smaller! energy bend less~more! and the electronswithin one radiation wavelength tend to get nearer to eother. This process produces microbunching of the electrat the scale ofl r . Electrons bunched within a wavelengemit coherent EM radiation, that is, the amplitude of the Efield is proportional to the number of electrons within thmicrobunch and the intensity proportional to the squarethe electrons. The larger intensity leads to more microbuning. The result is that the EM field keeps gaining enerfrom the bunched electrons and the radiation intensity (I r)grows exponentially withI r;e2z/LG, wherez is the undula-tor length. The gain length (LG) is defined asLG5lu /(4p)r), wherer, the dimensionless FEL parameter, is of the order of 0.001 or less.98 The radiation intensityeventually reaches saturation which occurs whenbunched electrons gain energy from the EM field balanby losing energy to the EM field. The process, called seamplified spontaneous emission,98,99 is illustrated in Fig. 45.To observe the SASE process, a high quality and high brig
FIG. 46. Exponential gain and saturation of SASE FELs at wavelength530 and 385 nm. Thex axis shows the integrated radiated energy and thyaxis the distance along the undulator system.~Reproduced with permissionfrom Ref. 100.!
FIG. 47. ~Color! Radiation damage to a lysozyme molecule as a functiontime. The simulated x-ray FEL intensity was 331012 photons~12 keV! per100 nm diam spot with the FWHM of the pulse 10 fs. The images showmolecule at the beginning, in the middle, and near the end of the x-ray p~Reproduced with permission from Ref. 103.!
e.n
3242 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
FIG. 48. Possible schematic layout for an experiment for imaging single biomolecules using x-ray FELs.
huanthala
nmn
m
raLsa
er-ter
enes.en-r-
by
ichls.ol-instoy
em-
ness electron beam is required. With the development of ptocathode radio-frequency electron guns and long, high qity undulators, SASE FELs operating at infrared, visible, aultraviolet wavelengths were experimentally observed inlate 1990s.100 More recently, a group at Argonne NationLaboratory has demonstrated the exponential gain and sration of SASE FELs at wavelengths of 530 and 385~shown in Fig. 46!,101 and a group at Deutsche ElektroneSynchrotron ~DESY!, operating the TESLA test facilitySASE FEL, has obtained exponential gain down to 80 nthe shortest wavelength obtained up to now for a FEL.102
B. Future prospects
The demonstration of SASE FELs at visible and ultviolet wavelengths paves the way for future x-ray FEbased on the SASE process. Tunable hard x-ray FELs h
o-l-
de
tu-
,
-
ve
already been proposed worldwide including the Linac Cohent Light Source at the Stanford Linear Accelerator Cenand TESLA at DESY.103,104 Due to the extremely high fluxand ultrashort pulses, x-ray FELs conceivably will opmany new opportunities in biology and biomedical sciencHere we illustrated one important application, i.e., the pottial for imaging single biomolecules using x-ray FELs. Curently, x-ray crystallography is the primary methodologywhich to determine the three-dimensional~3D! structure ofprotein molecules at near-atomic or atomic resolution, whrequires obtaining sizable good quality protein crystaHowever, somewhere around 20%–40% of protein mecules including most of the important membrane proteare difficult or impossible to crystallize. One possible wayovercome the crystallization difficulty is to extend x-racrystallography to noncrystals, which has recently been d
ity
FIG. 49. ~Color! ~a! One section of a 3D diffraction pattern processed from 106 identical copies of rubisco molecules with Poisson noise added and 33333 center pixel intensity removed. The edge of the diffraction pattern corresponds to 2.5 Å resolution.~b! Stereoview of the reconstructed 3D electron densmap of the rubisco molecule~contoured at 2s! on which an atomic model obtained from the Protein Data Bank is superimposed.~Reproduced with permissionfrom Ref. 110.!
onstrated by combining the coherent diffraction and the ovsampling phasing method.105 This novel approach can inprinciple be applied to imaging single biomolecules, but iposes very high radiation damage to biomolecules due toloss of crystallinity. With the prospects for x-ray FELs, thradiation damage problem may be circumvented. Theoresimulations show that, within about 10 fs, biomolecules cwithstand x-ray intensity of;3.83106 photons/Å2 withminimal structural change~as shown in Fig. 47!.106 In a com-bination of x-ray FELs and the novel approach of imaginoncrystalline specimens, a possible experimental setuoutlined here, shown in Fig. 48. X-ray FEL pulses will firbe focused downward to a 100 nm spot by a Fresnel zplate.107 Using a mass spectrometer, identical biomolecucan be selected and sprayed one by one in random orietion into the focused spot.108 Before scattering by a focusex-ray FEL pulse, each molecule will be orientated by a plarized nonresonant optical laser field.109 The diffraction pat-terns will be recorded by an x-ray CCD with fast readoThe experiment will be carried out in high vacuum to elimnate unwanted scattering. Two-dimensional~2D! diffractionpatterns from single molecules will be characterized andsembled into a 3D diffraction pattern. By employing thoversampling phasing method, it has been shown thasimulated 3D molecular diffraction pattern at 2.5 Å resotion can be successfully phased and transformed into ancurate electron density map comparable to that obtainedmore conventional methods~Fig. 49!.110 The powerful com-bination of the x-ray FEL and the new imaging approacould therefore have a tremendous impact on structural bogy.
XV. DISCUSSION
A wide range of biophysical and biomedical applicatioresearch has been accomplished by taking advantage ounique light source capabilities of FELs. These capabilitinclude combinations of wavelength ranges, pulse structuand peak and average power. A broad research and devment program resulted in human surgical applications usthe midinfrared Mark-III FEL. The operating parameters fan IR laser dedicated to surgical applications were specifiThe inhomogeneous buildup of heat that is key to controdegradation of material components seen in tissue ablaand mass spectrometry can confound spectroscopic meaments with infrared, linac-based FELs. A microbeam forvestigating tissue dynamics using multiple light sourcholds the promise of understanding at cellular length scai.e., a molecular description of cell physiology and biomecal processes.
The midinfrared SCA FEL and UV FELs based on stoage rings are particularly useful for one- and two-color sptroscopic investigations of biophysical processes. ThFELs enable spectroscopic techniques that access nwavelength and dynamic ranges not probed by complemtary techniques using conventional ultrafast lasers. Whilerepetition rates of these FELs are high enough to supnovel sampling techniques and keep the measurement t
r-
-he
aln
is
esta-
-
.
s-
a-c-
by
l-
thess,
op-g
rd.donre-
-ss,-
--e
veln-ertes
manageable, they are low enough to limit the confoundeffects of sample heating.
The THz-BRIDGE project currently supported by thEuropean Union includes the development of compact Ftechnology operating in the THz~far-infrared! range. Thegoals for applications of these THz sources include biolocal and diagnostic biomedical imaging.111
A source for pulsed, tunable, monochromatic x-rays pduced by inverse Compton scattering is now operationhaving benefited from a proof-of-principle demonstration uing the midinfrared Mark-III FEL. This source should enabnovel protocols in medical imaging as well as serve acompact source for studies of time-resolved structural bogy. Furthermore, the next generation SASE light soupromises extremely high flux and ultrashort pulses and cceivably will open many new opportunities in biological anbiomedical science.
1H. Motz et al., J. Appl. Phys.24, 826 ~1953!; J. M. J. Madey,ibid. 42,1906 ~1971!; V. L. Granatsteinet al., Appl. Phys. Lett.30, 384 ~1977!.
2C. A. Brau, Science239, 1115 ~1988!; Free-Electron Lasers~Academic,Boston, 1990!; H. P. Freund and T. M. Antonsen, Jr.,Principles of Free-Electron Lasers, 2nd ed.~Chapman & Hall, London, 1996!; P. O. Sheaand H. P. Freund, Science292, 1853~2001!.
3W. B. Colson, E. D. Johnson, M. J. Kelley, and H. A. Schwettman, PhToday 55, 35 ~2002! and references therein; National Research CounReport of the Committee on Free Electron Lasers~National Academy,Washington, DC, 1994!.
4R. Srinivasan, Science234, 559 ~1986!; IEEE J. Quantum Electron.QE-20, 1342 ~1984!; C. A. Puliafito, R. F. Steinert, and T. F. Duetsch, Opthalmology 92, 741 ~1985!; R. V. Jeffreys, inLasers in Neurosurgery,edited by R. V. Jeffreys~Wiley–Liss, 1992!, p. 79.
5R. R. Anderson and J. A. Parrish, Science220, 524 ~1983!; M. L. Wol-barsht, IEEE J. Quantum Electron.QE-20, 1427 ~1984!; D. Stern, C. A.Puliafito, E. T. Dobi, and W. T. Reidy, Ophthalmology95, 1434~1988!; J.T. Walsh, T. J. Flotte, and T. F. Deutsch, Lasers Surg. Med.9, 314~1989!;B. Zysset, J. G. Fujimoto, C. A. Puliafito, C. A. Birngruber, and T.Deutsch,ibid. 9, 193 ~1989!.
6C. A. Brau and M. H. Mendenhall, Nucl. Instrum. Methods Phys. Res331, ABS4 ~1993!.
7G. S. Edwards, B. Johnson, J. Kozub, J. Tribble, and K. Wagner, Opt. E~Bellingham! 32, 314 ~1993!.
8G. Edwardset al., Nature~London! 371, 416 ~1994!.9G. M. Peavy, L. Reinisch, J. T. Payne, and V. Venugopalan, Lasers SMed.25, 421~1999!; J. T. Payne, G. M. Peavy, L. Reinisch, and D. C. VaSickle, ibid. 29, 38 ~2001!.
10M. Ostertag, R. Walker, H. Weber, A. F. G. van der Meer, J. McKinley,Tolk, and B. Jean, Proc. SPIE2672, 181 ~1996!; E. J. Swift, Jr., G. S.Edwards, J. Perdigao, J. Y. Thompson, M. F. Nunes, D. E. Ruddell, anNegishi, J. Dent.29, 347 ~2001!.
11M. Speyer, J. Joe, J. M. Davidson, R. H. Ossoff, and L. Reinisch, Largoscope106, 845 ~1996!; B. Jean, Nucl. Instrum. Methods Phys. Res.393, 540~1997!; G. L. Bryant, Jr., C. F. Lano, Jr., L. Reinischet al., Ann.Otol. Rhinol. Laryngol.107, 269~1998!; G. L. Bryant, J. M. Davidson, R.H. Ossoffet al., Laryngoscope108, 13 ~1998!; D. S. Fortune, S. HuangG. L. Bryant et al., in Lasers in Surgery: Advanced CharacterizatioTherapeutics, and Systems VIII, edited by R. R. Anderson~1998!, BIOS/SPIE Vol. 3245, pp. 191–199; K. Awazu, A. Nagai, T. Tomimasuet al.,Nucl. Instrum. Methods Phys. Res. B144, 255 ~1998!; D. L. Ellis, N. K.Weisberg, J. S. Chenet al., Lasers Surg. Med.25, 1 ~1999!; S. R. Uhl-horn, S. Harrison, H. S. Pratistoet al., in Laser–Tissue Interaction X,edited by S. L. Jacques~1999!, BIOS/SPIE Vol. 3601, p. 3601.
12K. Becker, J. B. Johnson, and G. Edwards, Rev. Sci. Instrum.65, 1496~1994!.
13J. Tribble, D. Lamb, L. Reinisch, and G. Edwards, Phys. Rev. E55, 7385~1997!.
14J. Auerhammer, R. Walker, A. F. G. van der Meer, B. Jean, BookAbstracts, 18th FEL Conference and 3rd FEL User Workshop, Ro1996.
J.
h,
G
.
d-
rk
p
o
u
. D
m
-
D
lc
se
-
t.e,
ys
th-
es
,s.
.
. P.
H.
in
Rev.
se,
ec-
om.
ass
oc.
for
J.
i.
nce
J.d
m.
04
sasdu/
.ods
F.,.
em..
3244 Rev. Sci. Instrum., Vol. 74, No. 7, July 2003 Edwards et al.
15E. Sobol, A. Sviridov, M. Kitai, J. Gilligan, N. Tolk, and G. Edwards,Biomed. Opt.8, 216 ~2003!.
16G. S. Edwards, M. S. Hutson, S. Hauger, J. Kozub, J. Shen, C. ShieTopadze, and K. Joos, Proc. SPIE4633, 194 ~2002!.
17M. S. Hutson, S. A. Hauger, and G. Edwards, Phys. Rev. E65, 061906~2002!.
18D. C. Lamb, L. Reinisch, J. Tribble, R. H. Ossoff, T. J. Flotte, and A.Doukas, Proc. SPIE2391, 192 ~1995!.
19G. Edwards, Opt. Eng.~Bellingham! 34, 1524~1995!.20G. Edwardset al., IEEE J. Quantum Electron.QE-2, 810 ~1996!.21S. Grafstro¨m et al., Opt. Commun.65, 121 ~1988!.22H. S. Pratistoet al., in Specialty Fiber Optics for Medical Applications,
edited by A. Katzir and J. A. Harrington~1999!, BIOS/SPIE Vol. 3596, p.55.
23I. Gannot, A. Inberg, M. Oksmanet al., IEEE J. Quantum Electron.QE-2,880 ~1996!; I. Gannotet al., Appl. Opt.36, 6289~1997!.
24H. S. Pratisto, S. R. Uhlhorn, and E. D. Jansen, Fiber Integr. Opt.20, 83~2000!.
25Y. Matsuura, T. Abel, and J. A. Harrington, Appl. Opt.34, 6842~1995!; Y.Matsuuraet al., Electron. Lett.32, 1096~1996!.
26Y. Matsuura, K. Matsuura, and J. A. Harrington, Appl. Opt.35, 5395~1996!.
27G. P. Cram and M. L. Copeland, Nucl. Instrum. Methods Phys. Res144, 256 ~1998!.
28M. L. Copeland, R. J. Maciunas, and G. S. Edwards,Use of the Free-Electron Laser for Metastatic Brain Tumors in Neurosurgical Topics: Avanced Techniques in Central Nervous System Metastases, edited by R. J.Maciunas ~The American Association of Neurological Surgeons, PaRidge, II 1998!, p. 113.
29J. E. Bogganet al., Neurosurgery11, 609~1982!; J. E. Boggan, inLasersin Neurosurgery, Ref. 10, p. 21.
30J. H. Shen, J. A. Harrington, G. S. Edwards, and K. M. Joos, Appl. O40, 583 ~2001!.
31K. Matsuura, Y. Matsuura, and J. A. Harrington, Opt. Eng.~Bellingham! ,3418 ~1996!.
32E. A. J. Marcatili and R. A. Schmeltzer, Bell Syst. Tech. J.43, 1783~1964!.
33M. Miyagi and S. Karasawa, Appl. Opt.29, 367 ~1990!.34J. J. Corbet, J. A. Nerad, D. T. Tse, and R. L. Anderson, Arch. Ophthalm
~Chicago! 106, 1391~1988!.35K. M. Joos, J. H. Shen, D. J. Shetlar, and V. A. Casagrande, Lasers S
Med. 27, 191 ~2000!.36K. M. Joos, J. H. Shen, G. S. Edwards, D. J. Shetlar, J. Khoury, and R
Robinson, Invest. Ophthalmol. Visual Sci.37, S431~1996!.37D. R. Anderson, Trans. Am. Ophthalmol. Soc.79, 458 ~1981!.38W. Sun, J. H. Shen, D. J. Shetlar, and K. M. Joos, J. Glaucoma9, 325
~2000!.39C. A. Puliafito, R. F. Steinert, T. F. Deutsch, F. Hillenkamp, E. J. Deh
and C. M. Alder, Ophthalmology92, 741 ~1985!.40J. P. Cummings and J. T. Walsh, Jr., Appl. Phys. Lett.62, 1988~1993!.41B. A. Hooper, C. Lin, Y. Domankevitz, and R. R. Anderson,Therapeutic
Laser Applications, Technical Digest~Optical Society of America, Washington, DC, 1998!, p. 41.
42B. A. Hooper, Y. Domankevitz, C. Lin, and R. R. Anderson, Appl. Opt.38,5511 ~1999!.
43A. J. Welch, M. Motamedi, S. Rastegar, G. L. Le Carpentier, and E.Jansen, Photochem. Photobiol.53, 815 ~1991!.
44E. D. Jansen, T. G. van Leeuwen, M. Motamedi, C. Borst, and A. J. WeJ. Appl. Phys.78, 564 ~1995!.
45J. T. Walsh, Jr., T. H. Flotte, R. R. Anderson, and T. F. Deutsch, LaSurg. Med.8, 108 ~1988!.
46E. Hecht and A. Zajac,Optics~Addison–Wesley, Reading, MA, 1974!, p.81.
47F. Mirabella,Internal Reflection Spectroscopy, Practical Spectroscopy Series, Vol. 15~Dekker, New York, 1992!.
48R. Simhi, Y. Gotshal, D. Bunimovich, A. Sela, and A. Katzir, Appl. Op39, 3421~1996!; for near-field imaging with an IR FEL see, for examplM. K. Hong, A. Jeung, P. Huie, and S. Erramilli, Proc. SPIE3153, 115~1996!.
49H. D. Downing and D. Williams, J. Geophys. Res.80, 1656~1975!; G. M.Hale and M. R. Querry, Appl. Opt.12, 555 ~1973!.
50B. A. Hooper, C. Lin, Y. Domankevitz, and R. R. Anderson, Appl. PhLett. 78, 2381~2001!.
K.
.
B
t.
l.
rg.
.
,
.
h,
rs
.
51B. A. Hooper, G. C. LaVerde, and O. T. von Ramm, Nucl. Instrum. Meods Phys. Res. A475, 645 ~2001!.
52P. E. Young, A. M. Richman, A. S. Ketchum, and D. P. Kiehart, GenDev. 7, 29 ~1993!; F. Agnes and S. Noselli, C R. Acad. Sci. III322, 5~1999!; A. Jacinto, S. Woolner, and P. Martin, Devel. Cell3, 9 ~2002!; N.Harden, M. Ricos, Y. M. Ong, W. Chia, and L. Lim, J. Cell. Sci.112, 273~1999!; Q. J. Li, T. M. Pazdera, and J. S. Minden, Development~Cam-bridge, U.K.! 126, 2299~1999!; H. Liu, Y. C. Su, E. Becker, J. Treismanand E. Y. Skolnik, Curr. Biol.9, 101~1999!; Y. Lu and J. Settleman, GeneDev.13, 1168~1999!; M. G. Ricos, N. Harden, K. P. Sem, L. Lim, and WChia, J. Cell. Sci.112, 1225~1999!; K. Sawamotoet al., J. Cell Biol.146,361~1999!; V. Zecchini, K. Brennan, and A. Martinez-Arias, Curr. Biol.9,460 ~1999!.
53M. W. Bernset al., Science213, 507 ~1981!.54D. P. Kiehart, C. G. Galbraith, K. A. Edwards, W. L. Rickoll, and R. A
Montague, J. Cell Biol.149, 471 ~2000!.55M. S. Hutson, Y. Tokutake, M.-S. Chang, J. W. Bloor, S. Venakides, D
Kiehart, and G. S. Edwards, Science300, 145 ~2003!.56M. C. Halloran, M. Sato-Maeda, J. T. Warren, F. Y. Su, Z. Lele, P.
Krone, J. Y. Kuwada, and W. Shoji, Development~Cambridge, U.K.! 127,1953 ~2000!.
57A. E. Jeung, P. Huie, S. Erramilli, M. K. Hong, and H. A. Schwettmann,Ref. 14.
58Handbook of Biological Confocal Microscopy, edited by J. B. Pawley~Plenum, New York, 1990!.
59IMAGEJ, Ver. 1.24~available for free at http://rsb.info.nih.gov/ij/!.60For a broad overview of the state of the art, see for example, Chem.
~Washington, D.C.! 209–606, 101 ~2001!.61J. B. Fenn, M. Mann, C. K. Meng, S. F. Wong, and C. M. Whitehou
Science246, 64 ~1989!.62M. Karas, D. Bachmann, U. Bahr, and F. Hillenkamp, Int. J. Mass Sp
trom. Ion Processes78, 53 ~1987!.63R. Cramer, R. F. Haglund, Jr., and F. Hillenkamp, Int. J. Mass Spectr
Ion Processes169Õ170, 51 ~1997!.64M. R. Papantonakis, J. Kim, W. P. Hess, and R. F. Haglund, Jr., J. M
Spectrom.37, 639 ~2002!.65R. Cramer, K. Becker, R. F. Haglund, Jr., and F. Hillenkamp, J. Am. S
Mass Spectrom.7, 1187~1996!.66For a review of time-of-flight mass spectrometry instrumentation, see,
example, A. Vertes, R. Gijbels, and F. Adams,Laser Ionization MassAnalysis, Chemical Analysis series Vol. 124~Wiley, New York, 1993!.
67M. L. Baltz-Knorr, D. R. Ermer, K. E. Schriver, and R. F. Haglund, Jr.,Mass Spectrom.37, 254 ~2002!.
68M. L. Baltz-Knorr, K. E. Schriver, and R. F. Haglund, Jr., Appl. Surf. Sc197Õ198C, 19 ~2002!.
69See, for example, S. Berkenkamp, F. Kirpekar, and F. Hillenkamp, Scie281, 260 ~1998!.
70C. Menzel, K. Dreisewerd, S. Berkenkamp, and F. Hillenkamp, Int.Mass. Spectrom.207, 73 ~2001!; Z. Knochenmuss, R. Knochenmuss, anR. Knochenmuss, Mass Spectrom. Rev.17, 337 ~1998!.
71D. R. Ermer, M. Baltz-Knorr, and R. F. Haglund, Jr., J. Mass Spectro36, 538 ~2001!.
72A. Xie, A. van der Meer, W. Hoff, and R. H. Austin, Phys. Rev. Lett.84,5435 ~2000!.
73A. Xie, A. van der Meer, and R. H. Austin, Phys. Rev. Lett.88, 018102~2002!.
74D. Dlott and M. Fayer, J. Opt. Soc. Am. B6, 977~1989!; R. H. Austin, M.W. Roberson, and P. Mansky, Phys. Rev. Lett.62, 1912~1989!.
75R. H. Austin and A. Xie, Nucl. Instrum. Methods Phys. Res. A, 5~1998!.
76D. J. Segelstein, M.S. thesis, University of Missouri–Kansas City, KanCity, MO, 1981; a plot of these data can be found at http://omlc.ogi.espectra/water/gif/segelstein81.gif.
77M. E. Couprie, G. De Ninno, G. Moneron, D. Nutarelli, M. Hirsch, DGarzella, E. Renault, R. Roux, and C. Thomas, Nucl. Instrum. MethPhys. Res. A475, 617 ~2001!.
78M. E. Couprie, P. Tauc, F. Merola, A. Delboulbe´, D. Garzella, T. Hara, andM. Billardon, Rev. Sci. Instrum.65, 1485~1994!.
79E. Renault, L. Nahon, D. Nutarelli, D. Garzella, M. E. Couprie, andMerola, Proc. SPIE3925, 29 ~2000!; E. Renault, L. Nahon, D. GarzellaD. Nutarelli, G. De Ninno, M. Hirsch, and M. E. Couprie, Nucl. InstrumMethods Phys. Res. A475, 617 ~2001!.
80T. G. Scott, R. D. Spencer, N. J. Leonard, and G. Weber, J. Am. ChSoc. 92, 687 ~1970!; A. J. W. G. Visser and A. van Hoek, Photochem
Photobiol.33, 35 ~1981!; G. Krishnamoorthy, N. Periasamy, and B. Vekataraman, Biochem. Biophys. Res. Commun.144, 387 ~1987!; A. Gafniand L. Brand, Biochemistry15, 3165~1976!; D. W. Miles and D. W. Urry,J. Biol. Chem.243, 4181~1968!.
81O. Jardetzky and N. G. Wade-Jardetzky, J. Biol. Chem.241, 85 ~1966!.82S. L. Shapiro and K. R. Winn, J. Chem. Phys.73, 5958~1980!.83R. P. Haugland,Handbook of Fluorescent Probes~Molecular Probes, Eu-
gene, OR 1996!.84M. F. Goodman, Nature~London! 378, 237~1995!; A. Douhal, S. K. Kim,
and A. H. Zewail,ibid. 378, 260 ~1995!.85V. N. Litvinenko, S. H. Park, I. V. Pinayev, and Y. Wu, Nucl. Instrum
Methods Phys. Res. A470, 66 ~2001!.86M. S. Hutson, R. A. Palmer, A. Gillikin, M.-S. Chang, V. N. Litvinenko
and G. S. Edwards, Proc. SPIE4633, 225 ~2002!.87G. L. Carr, R. P. S. M. Lobo, C. J. Hirschmugl, J. LaVeigne, D. H. Reit
and D. B. Tanner, Proc. SPIE3153, 80 ~1997!; R. P. S. M. Lobo, J.LaVeigne, D. H. Reitze, D. B. Tanner, and G. L. Carr, Rev. Sci. Instru70, 2899~1999!.
88A. H. Lumpkin, B. X. Yang, V. Litvinenko, B. Burnham, S. Park, P. Wanand Y. Wu, Nucl. Instrum. Methods Phys. Res. A47, 338 ~1998!.
89K. Masutani, H. Sugisawa, A. Yokota, Y. Furukawa, and M. Tasumi, ApSpectrosc.46, 560 ~1992!.
90A. Sancar, Biochemistry33, 2 ~1994!.91T. Okamura, A. Sancar, P. F. Heelis, T. P. Begley, Y. Hirata, and N
Mataga, J. Am. Chem. Soc.113, 3143~1991!.92F. E. Carroll, J. W. Waters, R. H. Traegeret al., Proc. SPIE3614, 139
~1999!.93J. M. Boone, Radiology213, 23 ~1999!.94P. C. Johns and M. J. Yaffe, Phys. Med. Biol.32, 675~1987!; F. E. Carroll,
J. W. Waters, W. W. Andrewset al., Invest. Radiol.29, 266 ~1994!.95T. Takeda, A. Momose, and Y. Itaiet al., Acad. Radiol.2, 799 ~1995!.96D. Chapman, W. Thomlinson, R. E. Johnstonet al., Phys. Med. Biol.42,
2015 ~1997!.
,
.
.
.
97C. L. Gordon, Y. Yin, B. E. Lemofet al., Opt. Lett.20, 9 ~1995!.98R. Bonifacio, C. Pellegrini, and L. M. Narducci, Opt. Commun.50, 373
~1984!.99M. Kondratenko and E. L. Saldin, Sov. Phys. Dokl.24, 986 ~1979!.100M. J. Hoganet al., Phys. Rev. Lett.81, 4867~1998!; M. Babzienet al.,
Phys. Rev. E57, 6093 ~1998!; S. V. Milton et al., Phys. Rev. Lett.85,988 ~2000!.
101S. V. Milton et al., Science292, 2037~2001!.102J. Andruszkowet al., Phys. Rev. Lett.85, 3825~2000!.103M. Cornacchiaet al., Report SLAC-R-521, Stanford Linear Accelerato
Center, Stanford, CA~1998!.104DESY Report DESY97-048, edited by R. Brinkmann, G. Materlik,
Rossbach, and A. Wagner, Deutsches Elektronen Synchrotron, Ham~1997!.
105J. Miao, P. Charalambous, J. Kirz, and D. Sayre, Nature~London! 400,342 ~1999!.
106R. Neutze, R. Wouts, D. Spoel, E. Weckert, and J. Hajdu, Nature~Lon-don! 406, 752 ~2000!.
107R. Tatchyn, LCLS Optics: Technological Issues and Scientific Opponities in Proceedings of the Workshop on Scientific Applications of ShWavelength Coherent Light Sources, SLAC, Stanford, 1993.
108M. A. Tito, K. Tars, K. Valegard, J. Hajdu, and C. V. Robinson, J. AmChem. Soc.122, 3550~2000!.
109J. J. Larsen, K. Hald, N. Bjerre, H. Stapelfeldt, and T. Seideman, PRev. Lett.85, 2470~2000!.
110J. Miao, K. Hodgson, and D. Sayre, Proceedings of the National Acemy of Sciences 98, 6641~2001!.
111Proceedings of the THz-Bridge Workshop, Capri, 2002, J. Biol. Phys.~tobe published!.
112C. Abad-Zapatero, J. P. Griffith, J. L. Sussman, and M. G. RossmanMol. Biol. 198, 445 ~1987!, entry 1LDM of the Protein Data Bank.
