Selective absorption by melanin granules and selective cell targeting

Introduction

The use of short-pulsed lasers in ophthalmology dates back to the early 1960s. Soon after the development of the first ruby laser by Maiman,1 attempts were made to use this new light source for photocoagulation of the retina.2-4 However, the ruby laser did not always create reproducible lesions and was associated with a substantial risk of retinal hemor- rhage.4-7 When the argon ion laser became available about a decade later, it replaced the ruby laser as the instrument of choice for retinal photocoagulation because its wavelength was more favorable and because its continuous-wave (cw) output allowed better control of energy delivery. The pulsed ruby laser is no longer in use in ophthalmology today.

Several other ophthalmic lasers introduced in the 1970s and 1980s require short pulses to achieve their tissue effects, but they employ wavelengths that fall outside the visible spectrum, and do not require pigment absorption. Intraocular photodisruption using the Q-switched Nd:YAG laser,8,9 for example, is based on the principle of optical breakdown in transparent media, and relies on a nonlinear rather than a linear absorption mechanism. The fundamental output of Nd:YAG laser at 1064 nm is chosen to minimize absorption by the retina at this wavelength. Photoablation with the excimer laser at 193 nm, on the other hand, uses a wavelength that is strongly absorbed by the cornea and does not penetrate deeper into the eye.

In the last few years, there has been renewed interest in using short-pulsed lasers to target pigmented structures in the eye.13-17 Both selective laser trabeculoplasty (SLT, see the chapter by Latina in this volume) and selective RPE laser treatment (SRT, see the chapter by Roider, Brinkmann and Birn-

gruber) employ short pulses of green light, with melanin being the intended absorber. Moreover, both these new procedures aim to achieve cell-selective therapy, with the trabecular meshwork (TM) cells and the retinal pigment epithelial (RPE) cells as their respective targets, while minimizing collateral damage to adjacent tissue. Thus, the SLT targets the TM cells while preserving the underlying collagen beams of the meshwork; similarly, the SRT targets RPE cells while preserving the adjacent photorecep- tors.13-17

How can such selectivity be achieved? Given the problems encountered with the pulsed ruby laser in the earlier studies, how is it possible to make reproducible lesions in the TM and the RPE, and confine the damage to the target cells? What prevents gross disruption and creation of hemorrhage? This article focuses on the mechanism of short-pulse laser interaction with pigmented cells and with individual subcellular pigment granules. Better understanding of the cell damage mechanism is important, not only in the development of new laser targeting strategies, but also in setting safety standards for short-pulse laser exposures.18

Absorption of pulse laser radiation by melanin granules

The TM cells and RPE cells both contain numerous melanin granules, which are the primary absorbers for visible radiation in their respective parts of the eye.19-21 Energy deposition into the pigment absorbers results in a rise in tissue temperature; the spatial extent of the temperature increase depends on the duration of the laser pulse. Kapany et al. provided the following description on the impact of

ruby laser pulses on the retina, in one of the very early papers written on the subject:4 “During the time (~300 µsec) when light energy is delivered to the fundus, very little energy can escape to the surrounding tissue either by convection within the capillary system or by conduction to surrounding tissue. Because of this, the temperature of the small amount of tissue which absorbs the energy becomes highly elevated. If a large amount of energy is delivered rapidly, tissue fluid is vaporized and the vapor erupts through the retina into the vitreous humor and a bubble is formed. A hemorrhage or a retinal hole can also be caused under these conditions. If more moderate amounts of energy are delivered, the absorbed tissue are not vaporized and some of the stored energy is passed on to the retina causing it to be ‘welded’ to the choroid.”

In a well-known paper on selective photothermolysis, Anderson and Parrish gave a more in-depth account of the relationship between the optimum laser pulse duration and the size of the target to be treated.22 To confine the energy strictly to the ~1 µm melanin granules requires a pulse duration of less than 1 µsec. The 200-500-µsec pulse duration of the free running ruby laser, also known as normal mode or long-pulsed ruby laser (i.e., one that is not Q- switched), does indeed allow limited heat diffusion into the surrounding tissue (the distance of heat diffusion on this timescale is approximately 20 µm). Nevertheless, the description by Kapany et al. is essentially correct in that the pulsed ruby laser causes rapid, localized heating of a small amount of tissue, followed by vapor bubble formation leading to mechanical disruption if the local energy density exceeds a certain threshold value. Indeed, the ruby laser is not well-suited for creating thermal lesions (‘retinal weld’) because only a small volume of tissue is heated at low radiant exposures, whereas higher radiant exposures lead to vaporization, disruption, and hemorrhage.

With Q-switched (nanosecond) pulses, photomechanical tissue damage becomes even more pronounced. Marshall and Mellerio7 used terms such as ‘blast forces’ and ‘microexplosions’ to describe the histological appearance of retina lesions created by Q-switched ruby laser pulses. Damage always begins at the RPE, but they noted that, “Q-switched laser lesions, even those just above threshold, show a subretinal hemorrhage, while massive preretinal hemorrhage are obtained with only slightly more energy.” Thus, the range between threshold effect and serious hemorrhage was very limited.7

The granular nature of melanin absorption becomes less important when longer interaction times are considered (e.g., thermal effects created by the argon laser) because heat diffusion tends to smooth out any irregularity in the initial temperature distributions. Much of the thermal modelling work has been carried out by treating the RPE as a uniform absorbing layer, but more sophisticated models are available that take into account the subcellular distribution of the absorbing melanin granules.16,20,21,23

Microcavitation bubble formation around irradiated melanin granules

Ophthalmologists are familiar with the cavitation bubble formation that accompanies optical breakdown and plasma generation from the Q-switched Nd:YAG lasers.12 Typically these bubbles expand and collapse so rapidly that they are not visible with the unaided eye. The maximum expansion diameter for a bubble created by 1 mJ of pulse energy is about 1 mm, and the collapse time for such a bubble is about 100 µsec.12 Due to their short lifetime, the cavitation bubbles can only be detected with special techniques, such as high-speed (strobe) photography.12 Because optical breakdown requires a threshold of about 1 mJ for nanosecond pulses, the bubble size cannot be significantly smaller than ~1 mm. (Except under optimum focusing conditions using a high numerical aperture objective lens, bubble size of the order of 100 µm can be produced with microjoule pulse energies.)24

Now consider a 1-µm diameter spherical particle irradiated by a short laser pulse. Assume that the particle absorbs 63% of the photons that fall within the 1 µm circular cross-section of the sphere (this would require an absorption coefficient in excess of 10,000 cm-1). Then the total energy absorbed by the particle, for a radiant exposure of 55 mJ/cm2, is about 0.27 nJ, which is more than one million times smaller than the energy needed to create optical breakdown in the vitreous. Amazingly, a radiant exposure of 55 mJ/cm2 (nanosecond laser pulses at 532 nm) is sufficient to initiate bubble formation around the melanin particles. In other words, with a fraction of a nanojoule of energy, the particle is heated to a temperature high enough to vaporize fluid on its surface.

What are the sizes of these bubbles? Since the volume of a bubble is to a first approximation proportional to its energy E, the diameter of the bubble is proportional to E-1/3.12 Compared to a typical bubble induced by optical breakdown (1 mm at 1 mJ), a bubble created from melanin granular absorption of 1/106 the energy should have a diameter that is 102 times smaller, or about 10 µm in size. The lifetime of a 10-µm bubble is estimated to be about 1 µsec based on the Rayleigh formula.25

Experimental evidence of bubble formation around laser-irradiated melanin granules was obtained26 using a microscope capable of very high-speed image capture with nanosecond time resolution (Fig. 1).

The experiment was done as follows:26 A microscope was set up to image individual melanosomes isolated from fresh bovine RPE cells. A laser pulse was split into two pulses. The first one (532 nm) was delivered through the microscope to irradiate the particles. The second pulse was used to generate strobe light (~565 nm) that illuminates the sample at various delay times after the first pulse to produce a high-speed ‘stop-action’ image. The arrival time of the strobe pulse was determined by the length of

Fig. 1. Nanosecond time-resolved imaging of microcavitation bubble formation around individual laser-heated melanosomes.

an optical delay line and was variable from <1 nsec to >160 nsec. This set-up eliminates timing jitter, and the time resolution of this imaging system is determined by the laser pulse duration – nanoseconds with Q-switched lasers and picoseconds with modelocked lasers.

In a typical experiment, three images were taken in sequence, as shown in Figure 1. The first image on the left was taken before laser irradiation; three melanosomes were visible in this particular example. The second image (middle) was taken a few nanoseconds after the particles were heated by a 532 nm laser pulse. An expanding bubble can clearly be seen around each particle. Bubbles appear dark in the bright field (trans-illumination) image because the strobe light is scattered out of the collection optics of the microscope by the bubbles. The third image on the right was taken after the bubbles have collapsed, showing apparently intact particles.

Bubble dynamics around a single particle were investigated by measuring the intensity of a lowpower helium-neon laser probe beam, focused to a small spot around the microparticle.26 As the bubble grows, the intensity of the forward-transmitted beam is attenuated, while the backscattered signal increases. By using a microscope with three different laser colors (532 nm for irradiation, 565 nm for strobe light, and 633 nm for the probe beam), we are able to create and image the cavitation bubble and detect the bubble dynamics simultaneously. Figure 2

shows the detected signal in the forward (transmitted) direction when a single melanosome particle was irradiated. At radiant exposure level below cavitation threshold, no signal was detected. (The upward spike at the beginning of each trace was due to leakage of the strong pump pulse radiation into the photodiode detector.) For radiant exposures above cavitation threshold, a transient decrease in the transmitted intensity of the probe beam was observed as the result of attenuation by the bubble. The lifetimes of the bubble obtained from the lower two traces shown in Figure 3 were 580 nsec at 1.4 × threshold and 850 nsec at 2.2 × threshold. Corresponding bubble diameters, obtained from images taken with 125nsec strobe delay, were 5.5 and 7.5 µm, respectively.

Both the bubble diameter and the bubble lifetime increased with increasing radiant exposure. The measured values were in the range of 1-10 µm (diameter) and 0.1-1 µsec (lifetime) for radiant exposures that ranged from just above threshold to a few times above threshold. Under these conditions, the particles appeared to stay intact after bubble collapse. A single particle can be irradiated repeatedly, producing a bubble each time, without being destroyed. This observation is consistent with the notion that the bubble comes from fluid vaporization at the surface of the particle (a ‘vapor blanket’) rather than vaporization of the particle itself.27 A bubble that contains hot vapor inside and surrounded by cooler fluid outside is an unstable bubble. The vapor rapidly condenses, leaving an ‘empty’ cavity with low internal pressure, and the bubble must collapse. This is the reason for the short bubble lifetime. The dynamics of bubble expansion and implosion is similar to the cavitation dynamics previously observed with laser-induced breakdown, except on a much smaller scale. This is a good example of cavitation by local energy deposition as defined by Lauterborn.28

The threshold for bubble formation around bovine melanosomes is equal to 55 mJ/cm2 for nanosecond and shorter laser pulses at 532 nm.26 In this regime, the threshold is independent of the pulse duration as long as the pulse duration is shorter than the thermal relaxation time of the particles. As pulse duration increases, becoming comparable then exceeding the thermal relaxation time, the threshold increases, from ~300 mJ/cm2 for 1-µsec pulses, to ~400 mJ/cm2 for 2-µsec pulses, and ~550 mJ/cm2 for 3-µsec pulses.29 For these longer pulse durations, more energy escapes from the particle into the surrounding fluid. Therefore, higher radiant exposures are required to reach the temperature for vaporization.

What is the temperature for vaporization at the surface of the melanin particle? Does this happen at 100°C, as with normal boiling of water? The answer is no, because bubble formation requires nucleation. In the absence of nucleation, liquid water can be superheated well beyond 100°C without turning into vapor. The reason is surface tension: excess energy is needed to overcome surface tension when

creating a new interface in the fluid. Pre-existing bubbles or impurities (nucleation centers) lower this energy requirement by allowing bubbles to grow from the existing surface. The pressure Psur associated with the surface tension σ is inversely proportional to the radius of the surface r according to

Psur=2*σ/r.30 In the case of melanin in water, the particle itself is acting both as a heat source and as

a nucleation site. For a bubble to grow from an initial diameter of 1 µm (diameter of the nucleation center), it has to overcome a total pressure of about 3.5 atm (the hydrostatic pressure plus Psur). The boiling point of water increases with increasing pressure. At 3.5 atm the vaporization temperature is 137°C. Experimentally the particle temperature has been determined to be about 150°C.29

Intracellular cavitation bubble formation and selective cell killing

The preceding section describes bubble formation around isolated melanin particles. What happens

when they are produced in cells that contain these particles? Using cultured TM cells, Latina and Park31 first showed that nanosecond laser irradiation selectively induce lethality in cells containing pigment particles. Is bubble formation the mechanism for cell killing? To answer this question, we performed highspeed imaging of laser-irradiated TM cells to visualize bubble formation in these cells.32 We also added an argon ion laser for fluorescence excitation and an integrating CCD camera for fluorescence detection. This set-up allowed us to irradiate cells, image bubble formation (on the nanosecond timescale), and perform fluorescence cell viability assay all under the same microscope without moving the cells. With high-speed imaging, we observed intracellular microbubble formation around melanin particles that have been ingested by the TM cells. Moreover, in any given cell when bubble was created, the cell lost

viability. Conversely, when bubble was not produced, the cell survived.32

Selective cell killing

Selective cell killing was investigated using mixed TM cell culture, as described by Latina and Park.31 Pigmented and nonpigmented TM cells were grown separately, then mixed together before the experiment. Irradiation of the mixed population at radiant exposure above bubble formation threshold results in killing of every cell that contain melanin, while nearby nonpigmented cells remain viable (Fig. 4).

Similar experiments were performed with primary RPE cells using tissue explants from freshly enucleated bovine eyes. Bubble formation in the RPE results in cell death (Fig. 5). For nanosecond and shorter pulses, the threshold for RPE cell death is 55 mJ/

cm2, identical to the threshold for bubble formation around single melanosomes.33

Is this the same kind of bubbles as those implicated in the ruby laser induced retinal hemorrhage in the earlier studies? While Kapany et al. talked about vapor eruption through the retina and into the vitreous humor,4 our microbubbles expand and collapse entirely within the confines of the RPE or TM cells. These short-lived (microsecond) microbubbles are not visible by ophthalmoscopy and were most likely not noticed in the previous studies. In fact, Roider et al.34 investigated bubble formation as a damage mechanism for 200-nsec laser pulses, using ophthalmoscopically visible bubbles as an endpoint. They found the threshold for what they called macroscopic bubble formation (those visible by ophthalmoscopy) to be about ten times higher than the angiographic threshold for retinal damage; the threshold for hemorrhage was even higher. Histological examination of the lesions irradiated at the angiographic threshold show damage located primarily to the RPE, which were often physically separated from the Bruch’s membrane and from the photoreceptor outer segments.34 The radiant exposure at angiographic threshold was 45 mJ/cm2 (ten pulses at 532 nm, 200 nsec pulse duration), a value that is comparable to our measured threshold (55 mJ/cm2) for bubble formation around single melanin granules and for RPE cell death ex vivo. (It should be cautioned, however, that accurate comparison of radiant exposures is very difficult due to the uncertainty in determining the spot size accurately in vivo.) We conclude that the mechanism for RPE damage at the angiographic threshold level is microbubble expansion and collapse within the RPE cells. At this level, neither the lesion nor the bubbles themselves are ophthalmoscopically visible. When the radiant exposure was increased to about ten times the threshold for cavitation, stable gas bubbles were observed after the cavitation bubble collapse. These gas bubbles are the product of ‘rectified diffusion’ that occurs when the lifetime of the cavitation bubble becomes sufficiently long to allow dissolved gas to escape from the fluid and diffuse into the cavity. These stable bubbles are visible by ophthalmoscopy and are what Roider et al. identified as macroscopic bubbles.34

It is clear from the above discussion that different authors mean different things when using the term bubble formation, a term that has appeared in many

papers on retinal damage. Readers are advised to check the definition and not to assume that all bubbles are the same.

Implications for selective laser trabeculoplasty and selective retinal pigment epithelium laser treatment

Perhaps the most interesting finding from these studies is the observation that cavitation damage (a kind of microscopic underwater explosion) can be confined to the scale of single cells. Precise localization is possible because of the minute amount energy absorbed by each melanin granule (a fraction of a nanojoule), producing cavitation bubbles that are only a few micrometers in size. Thus, even though the damage process is photomechanical in nature, selective cell targeting can be achieved using short pulsed lasers without causing gross tissue disruption.

In principle, the size of the bubble should be ‘selflimiting’, meaning that once a bubble is created, it forms an insulating blanket around the particle (the thermal conductivity for water vapor being much lower than that for liquid water). The vapor blanket should both scatter incoming radiation and limit further heat transfer from the particle to the surrounding fluid. In practice, we have not seen evidence of such a self-limiting process. The bubble size always increased with increasing pulse energy. Consequently, proper setting of laser energy is essential for ensuring that the target cells are successfully treated without creating excess damage to collateral tissue. For SLT, unlike conventional argon laser trabeculoplasty (ALT), there is no visible endpoint. To find the proper energy setting, the current protocol for SLT calls for gradual increase of pulse energy until bubble formation is observed, then backing down until the energy is below the bubble formation threshold.36 What do they mean by bubble formation? This is an example why caution is warranted when encountering the term bubble formation. Surely what they mean here is the stable (gas) bubble formation, because conventional slit-lamp illumination is being used to observe the anterior chamber in this procedure. When the pulse energy is reduced below bubble formation threshold, as instructed by the protocol, we are in the regime of the transient in-

tracellular cavitation bubble formation that is not visible by ophthalmoscopy. This is the proper treatment setting for SLT. An automatic method to monitor the transient cavitation bubble formation can make this process more reliable and less time consuming. It remains an interesting open question whether any biological response can be elicited if the radiant exposure is further reduced to below the threshold for transient intracellular bubble formation (for example, intracellular hot spots can be produced by heating the melanin granules to 120°C without bubble nucleation).

For SRT, the requirement for selectivity is probably more stringent than for SLT if we aim to preserve the photoreceptors adjacent to the treated RPE cells. Microsecond pulses were originally chosen15,16 precisely to avoid photomechanical damage, but evidence now points to the possibility of microbubble formation inside the RPE as the potential damage mechanism even for pulse duration lasting a few microseconds.29 If so, monitoring bubble formation can be used as an online control for laser setting to ensure that the damage is confined to the RPE cells. Interesting remaining scientific questions are: (a) at what pulse duration does the crossover from photomechanical to photothermal damage takes place; and (b) what pulse duration offers the best therapeutic bandwidth, defined as the ratio of the threshold for photoreceptor damage to the threshold for selective RPE damage? Resolution of these questions will enable ophthalmologists to answer the most important medical question, that is, the true value of SRT in treating RPE-related disorders, particularly in macular degeneration.

Conclusions

A very small amount of energy (less than 1 nJ) deposited in a very small volume (a single melanosome) can produce very high local temperature, which in turn causes a very small underwater explosion (microcavitation) with damage range on the scale of single cells. The subtle damage caused by the microcavitation bubbles are not visible under ophthalmoscopic examination. Had the existance of these transient microbubbles been recognized, the development of cell-selective laser therapies such as SLT and SRT could have come much earlier, perhaps even before the introduction of cw argon laser photocoagulation, using the pulse ruby laser.

Acknowledgments

I wish to thank Michael W. Kelly, whose PhD thesis research resulted in many of the results presented here, Dr Santiago Sibayan for providing cultured TM cells, and Jan Roegener, Clemens Alt, Ralf Brinkmann, Gereon Huettmann, Reginald Birngruber, Johanne Roider, Franz Hillenkamp, Mark Latina, Rox Anderson, and David Sliney for stimulating discussions. This work was supported by NIH EY12970 and AFOSR F49620- 00-1-0179.

References

1.Maiman TH: Stimulated optical radiation in ruby. Nature 187:493-494, 1960

2.Zaret et al: Ocular lesions produced by an optical maser (laser). Science 134:1525, 1961

3.Cambell CJ, Rittler MC, Koester CJ: The optical maser as a retinal coagulator: an evaluation. Trans Am Acad Ophthalmol 67:58, 1963

4.Kapany NS, Peppers NA, Zweng HC, Flocks M: Retinal photocoagulation by lasers. Nature 199:146-149, 1963

5.Marshall J, Mellerio J: Histology of the formation of retinal laser lesions. Exp Eye Res 6:4-9, 1967

6.Marshall J, Mellerio J: Pathological development of retinal laser photocoagulations. Exp Eye Res 6:303-308, 1967

7.Marshall J, Mellerio J: Histology of retinal lesions produced with Q-switched lasers. Exp Eye Res 7:225-230, 1968

8.Fankhauser F, Roussel P, Steffen J, Van der Zypen E, Chrenkova A: Clinical studies on the efficiency of high power laser radiation upon some structures of the anterior segment of the eye: first experiences of the treatment of some pathological conditions of the anterior segment of the human eye by means of a Q-switched laser system. Int Ophthalmol 3:129-139, 1981

9.Fankhauser F, Lortscher H, Van der Zypen E. Clinical studies on high and low power laser radiation upon some structures of the anterior and posterior segments of the eye: experiences in the treatment of some pathological conditions of the anterior and posterior segments of the human eye by means of a Nd:YAG laser, driven at various power levels. Int Ophthalmol 5:15-32, 1982

10.Loertscher HP: Laser-induced breakdown for ophthalmic applications. In: Trokel SL (ed) YAG Laser Ophthalmic Microsurgery, pp 39-66. Norwalk, CN: Appleton-Century- Crofts 1983

11.Docchio F, Dossi L, Sacchi CA: Q-switched Nd:YAG laser irradiation of the eye and related phenomena: an experimental study. I. Optical breakdown determination for liquids and membranes. Lasers Life Sci 1:87-103, 1986

12.Vogel A, Schweiger R, Frieser A, Asiyo M, Birngruber R: Intraocular Nd:YAG laser surgery: light-tissue interaction, damage range, and reduction of collateral effects. IEEE J Quantem Electr QE-26:2240-2260, 1990

13.Latina MA, Sibayan SA, Shin DH, Noecker RJ, Marcellino

G:Q-switched 532-nm Nd:YAG laser trabeculoplasty (selective laser trabeculoplasty): a multicenter, pilot, clinical study. Ophthalmology. 105:2082-2088, 1998

14.Park CH, Latina MA, Schuman JS: Developments in laser trabeculoplasty. Ophthalmic Surg Lasers 31:315-322, 2000

15.Roider J, Michaud NA, Flotte TJ, Birngruber R: Response of the retinal pigment epithelium to selective photocoagulation. Arch Ophthalmol 110:1786-1792, 1992

16.Roider J, Hillenkamp F, Flotte TJ, Birngruber R: Microphotocoagulation: selective effects of repetitive short laser pulses. Proc Nat Acad Sci US 90:8463-8647, 1993

17.Roider J, Brinkmann R, Wirbelauer C, Laqua H, Birngruber

R:Retinal sparing by selective retinal pigment epithelial photocoagulation. Arch Ophthalmol 117:1028-1034, 1999

18.Sliney DH, Marshall J: Tissue specific damage to the retinal pigment epithelium: mechanisms and therapeutic implications. Lasers Light Ophthalmol 5:17-28, 1992

19.Wolbarsht ML, Fligsten KE, Hayes JR: Retina: pathology of neodymium and ruby laser burns. Science 150:1453-1454, 1965

20.Hayes JR, Wolbarsht ML: Thermal model for retinal damage induced by pulsed lasers. Aerospace Med 39:474-480, 1968

21.Hansen WP, Fine S: Melanin granule models for pulsed laser induced retinal injury. Appl Opt 7:155-159, 1968

22.Anderson RR, Parrish JA: Selective laser photothermolysis: precise microsurgery by selective absorption of laser radiation. Science 220:524-527, 1983

23.Thompson CR, Gerstman BS, Jacques SL Rogers ME: Melanin granule model for laser-induced thermal damage in the retina. Bull Math Biol 58:513-553, 1996

24.Venugopalan V, Guerra A 3rd, Nahen K, Vogel A: Role of laser-induced plasma formation in pulsed cellular microsurgery and micromanipulation. Phys Rev Lett 88:78-103, 2002

25.Rayleigh L: On the pressure developed in a liquid during the collapse of a spherical cavity. Phil Mag 34:94-98, 1917

26.Lin CP, Kelly MW: Cavitation and acoustic emission around laser-heated microparticles. Appl Phys Lett 72:2800, 1998

27.Pustovalov VK: Thermal processes under the action of laser radiation pulse on absorbing granules in heterogeneous biotissues. Int J Heat Mass Transfer 36:391-399, 1993

28.Lauterborn W: Cavitation and Inhomogeneities in Underwater Acoustics. Springer-Verlag 1980

29.Brinkmann R, Hüttmann G, Rögener J, Roider J, Birngruber R, Lin CP: Origin of RPE-cell damage by pulsed laser irradiance in the ns to µs time regime. Lasers Surg Med 27:451–464, 2000

30.Young F: Cavitation. McGraw Hill 1989

31.Latina MA, Park C: Selective targeting of trabecular meshwork cells: in vitro studies of pulsed and CW laser interactions. Exp Eye Res 60:359-371, 1995

32.Lin CP, Kelly MW, Sibayan SA, Latina MA, Anderson RR: Selective cell killing by microparticle absorption of pulsed laser radiation. IEEE J Sel Top Quant Electron 5:963-968, 1999

33.Kelly WM, Lin CP: Microcavitation and cell injury in RPE cells following short-pulsed laser irradiation. Proc SPIE 2975:174-179, 1997

34.Roider J, El Hifnawi E, Birngruber R: Bubble formation as the primary interaction mechanism in retinal laser exposure with 200-nsec laser pulses. Lasers Surg Med 22:240248, 1998

35.Gerstman BS, Thompson CR, Jacques SL, Rogers ME: Laser Induced Bubble Formation in the Retina. Lasers Surg Med 18:10-21, 1996

36.Latina MA, Tumbocon JA: Selective laser trabeculoplasty: a new treatment option for open angle glaucoma. Curr Opin Ophthalmol 13:94-96, 2002

Introduction

In the late 1970s, Krasnov introduced the use of focused Q-switched ruby laser pulses for goniopuncture and iridotomy in the treatment of glaucoma,1 and, in 1980, Aron-Rosa et al.2 reported performing posterior capsulotomies following extracapsular cataract surgery employing picosecond (psec) pulse trains generated by a mode-locked Nd:YAG laser. Shortly afterward, following extensive preliminary morphological studies,3 Fankhauser et al. published their first reports on laser surgery in the anterior and posterior sectors of the eye,4,5 which were conducted using pulses from a Q-switched Nd:YAG laser (pulse duration, ≈10 nsec). The possibility of operating surgically by means of ‘photodisruption’ on pigmented and nonpigmented structures without opening the eye aroused great enthusiasm among ophthalmologists,6-9 as well as concerns about safety because of the explosive character of the laser effects and the fact that part of the laser light is transmitted through the laser plasma onto the retina.10,11 For this reason, a vigorous debate developed about the advantages and disadvantages of Q-switched lasers as opposed to mode locked lasers, which produce pulse trains of 7-10 psec pulses separated by 5-8 nanosecond (nsec) intervals. This debate initially relied primarily on the available literature on fundamental physical effects,6,10,12,13 but specific studies were soon undertaken on the thresholds for optical breakdown,14-16 plasma transmission,14,17-20 and mechanical laser effects. The major inadequacy of many of these studies was that they only dealt with individual aspects of the complex lasertissue interaction. Later, a more consistent picture evolved on the basis of systematic studies that examined the complete sequence of different physical

mechanisms involved in intraocular photodisruption and ablation (plasma formation, shock wave production, cavitation), and the resulting tissue effects. The present article aims at presenting a concise but comprehensive portrait of this picture. Thereby, special emphasis is laid on explaining the physical basis for the fascinating new possibilities that emerged with the availability of compact femtosecond (fsec) laser sources as, for example, intrastromal corneal refractive surgery,27-29 and intracellular surgery.30,31

Kinetics of plasma formation in biological tissues

Laser-induced, plasma-mediated ablation, also known as laser-induced breakdown, relies on nonlinear absorption in the target, which is achieved when a ma- terial-specific irradiance threshold is exceeded.32-34 In tissues with strong linear absorption, plasma formation can be initiated by thermionic emission of (quasi-)free electrons. In this case, the plasma formation usually ‘shields’ the underlying structures and impedes further energy deposition by linear absorption.35 However, plasma can also be formed in materials that are transparent at low irradiance. This occurs at high irradiance when seed electrons for an ionization avalanche are provided by multiphoton ionization. The process then progresses through an interplay of multiphoton ionization and avalanche ionization of target molecules. Thus, plasma formation provides a unique possibility for the achievement of a highly-localized energy deposition in transparent or low-absorbing materials. Localized energy deposition is achieved by using focused laser radiation because plasma formation is limited to regions where the irradiance is high enough to exceed the threshold for laser-induced breakdown.

Experimental studies have demonstrated that the optical breakdown threshold in water is similar to that in transparent ocular and other biological media (cornea, vitreous fluid, and saline).15,35 For convenience, we shall therefore focus our attention on the physics of plasma formation in pure water. While optical breakdown in gases leads to the generation of free electrons and ions, electrons in liquids are either bound to a particular molecule or are ‘quasifree’ when they have sufficient kinetic energy to move without being captured by local molecular energy potentials. Thus, transitions between bound and quasi-free states are the equivalent of ionization of molecules in gases. In order to describe the breakdown process in water, we adopt Sacchi’s approach, who proposed to treat water as an amorphous semiconductor with an excitation energy of ∆E = 6.5 eV, corresponding to the transition from the molecular 1b1 orbital into an excitation band.37,38 For simplicity, we use the terms ‘free electrons’ and ‘ionization’ as abbreviations for ‘quasi-free electrons’ and ‘excitation into the conduction band’.

The process of plasma formation is schematically depicted in Figure 1. It essentially consists of the formation of free electrons by an interplay of multiphoton ionization and avalanche ionization. The promotion of an electron from the ground state to the valence band requires the energy of two photons for UV wavelengths of up to λ = 383 nm, three photons for wavelengths of up to 574 nm, and four, five, and six photons for wavelengths of up to 766, 958, and 1153 nm, respectively. In pure water, this energy can only be provided when several photons interact simultaneously with a bound electron. The multiphoton ionization rate is proportional to Ik, where I is the laser light irradiance and k the number of photons required for ionization.

Once a free electron exists in the medium, it can

absorb photons in a non-resonant process called ‘inverse Bremsstrahlung absorption’ (IBA) in the course of collisions with heavy charged particles (ions or atomic nuclei).39 A third particle (ion/atom) is necessary for conserving energy and momentum during optical absorption. Absorption of the photon increases the kinetic energy of the free electron. After k IBA events, the kinetic energy of the electron exceeds the band gap energy ∆E, and the electron can produce another free electron via impact ionization. After impact ionization, two free electrons with low kinetic energies are available which can again gain energy through IBA. The recurring sequences of IBA events and subsequent impact ionization lead to a rapid growth in the number of free electrons, if the irradiance is sufficient to overcome the losses of free electrons through diffusion out of the focal volume and through recombination. In addition, the energy gain through IBA must be more rapid than energy losses through collisions with heavy particles. Energy is lost because a fraction of the kinetic energy of the electron that is proportional to the ratio of the electron and ion masses is transferred to the ion during each collision. The process involving both IBA and impact ionization is called ‘avalanche’ or ‘cascade’ ionization. At high irradiances, the losses play a minor role, and the cascade ionization rate for a given number density of free electrons is proportional to the irradiance.32

Multiphoton ionization occurs on a time scale of a few femtoseconds, and the multiphoton ionization rate is independent of the number density of free electrons. In contrast, cascade ionization depends on the number density of free electrons at the laser focus and requires a longer time because several consecutive IBA events are necessary for a free electron to acquire the kinetic energy for impact ionization. With an ionization energy of 6.5 eV and a photon energy of, for example, 1.56 eV (corresponding to

λ = 800 nm), an electron must undergo at least five IBA events before it can produce another free electron through impact ionization. As mentioned above, IBA can only occur during collisions of the electrons with heavy particles. In condensed matter, the time τ between collisions is roughly 1 fsec.40 Thus, even at extremely high irradiance where most collisions involve IBA, every doubling of the number of free electrons requires at least 5 fsec. Due to this time constraint, avalanche ionization can contribute significantly to plasma formation for laser pulse durations in the fsec range only after a large number density of free electrons has been provided by multiphoton ionization.

Several authors have used rate equations based on the Drude model to describe the temporal evolution of the volumetric density of free-electrons ρ under the influence of the laser radiation and to calculate breakdown thresholds for various laser parameters. The generic form of such a rate equation is

The first two terms on the right hand side of Equation (1) represent the production of free electrons through multiphoton and cascade ionization, respectively. The last two terms describe losses through diffusion of electrons out of the focal volume and recombination. The cascade ionization rate ηcasc and the diffusion loss rate g are proportional to the density of free electrons, g, while the recombination rate ηrec is proportional to ρ 2, as it involves an interaction between two charged particles (an electron-hole pair). A detailed description of the individual terms of Equation (1) has been given by Kennedy41 and Noack and Vogel.46

Several investigations, based on the above rate equation, neglected either multiphoton ionization,32,42 recombination,41,45 or diffusion,44 and all four terms of the rate equation have only been considered in a few publications.43,46,47 While many of the early studies were focused on calculation of the breakdown thresholds, recent numerical simulations also included an analysis of the time evolution of the electron density during the laser pulse, the irradiance dependence of the free-electron density, plasma absorption, and volumetric energy density in the plasma.46,47

The description of the plasma formation process is complicated by the fact that both the refractive index and absorption coefficient depend on irradiance. This leads to a spatial phase modulation of the wave front of the laser beam that depends on the intensity distribution across the beam. This modulation brings about a change in the intensity distribution (self-focusing or defocusing) that, in turn, has an effect on the nonlinear absorption process. The degree of self-focusing increases with decreasing focusing angle and shorter laser pulse dura-

tion. Self-focusing is more prominent in these conditions as it requires a critical power to be exceeded that is largely independent of the focusing angle or pulse duration used.34 In contrast, optical breakdown requires an irradiance threshold to be surpassed. The power necessary to provide this irradiance becomes larger with larger spot size (i.e., decreasing focusing angle) and decreasing laser pulse duration. Therefore, the optical breakdown threshold is at sufficiently small focusing angles and at short pulse durations larger than the critical power for self-focusing, and the breakdown process is influenced by self-focusing, leading to the formation of plasma filaments.48

Threshold for plasma formation

The threshold radiant exposure for breakdown determines the minimum achievable extent of the laser effect used for laser ablation or dissection. From an experimental perspective, the threshold for nsec and psec laser-induced breakdown in aqueous media is defined by the irradiance or radiant exposure leading to the observation of a luminescent plasma at the laser focus.51 With shorter laser pulses, there is no plasma luminescence in the visible region of the spectrum, and breakdown is experimentally detected by the observation of a cavitation bubble in the liquid.46,52 From a theoretical point of view, optical breakdown is identified by the generation of a critical free-electron density between ρcr = 1018 cm-3 and 1021 cm-3.41,43,46,51 A good match between experimental threshold values and theoretical predictions for optical breakdown in water is obtained when critical electron densities of ρcr = 1020 cm-3 for nsec pulses and ρcr = 1021 cm-3 for psec and fsec pulses are assumed.46

The irradiance threshold for plasma generation increases by three orders of magnitude when the laser pulse duration is decreased by six orders of magnitude from the nsec range into the fsec range, as shown in Figure 2. The increase in irradiance is required to compensate for the reduced time available to reach the critical electron density. Remarkably, the threshold radiant exposure decreases by three orders of magnitude for the same decrease of pulse duration. Two reasons are responsible for this decrease: (a) the proportionality of the multiphoton ionization rate with Ik, and (b) the decrease in the plasma energy density with shorter laser pulse durations that is explained further below.

For pulse durations in the nsec and psec range, the plasma formation thresholds are considerably reduced when the target has a high linear absorption coefficient, because the seed electrons for avalanche ionization are provided by thermionic emission of free electrons. The threshold for plasma formation in transparent media (water or cornea) using pulses of a few nsec is of the order 100-400 J/cm2.48,53 In contrast, plasma formation for ArF excimer laser

Fig. 2. Threshold irradiance and radiant exposure for optical breakdown in water versus laser pulse duration. The data are from Noack and Vogel46 and Vogel et al.70 Note that all experimental data were obtained for a focusing angle of around 20° with almost diffraction-limited focusing. Aberrations in the optical delivery system lead to severe errors in the threshold values, even if the focal diameter is determined experimentally.67

Fig. 3. a. Temporal evolution of free electron density during laser irradiation, for 1064-nm wavelength and 6-nsec pulse duration. The time t is normalized with respect to the pulse duration tp. The contribution of multiphoton ionization to the total free-electron density is plotted as a dotted line. b. Irradiance dependence of the maximal free electron density ρmax for the same laser parameters. The irradiance is normalized with respect to the calculated threshold irradiance Irate. The thresh-

old Irate and the corresponding value of ρmax are marked with dotted lines. (Reproduced from Vogel et al.47 by courtesy of

the publisher.)

Fig. 4. a. Temporal evolution of free electron density during laser irradiation and b. irradiance dependence of the maximal free electron density for λ = 532 nm and tp = 100 fsec. (Reproduced from Vogel et al.47 by courtesy of the publisher.)

ablation of skin (λ = 193 nm, tL = 22 nsec) where the linear absorption is very high (µa ≈ 40,000 cm-1),54 was reported for radiant exposures of as small as 0.25 J/cm2,55 and in TEA CO2 laser ablation of skin (λ = 10.6 µm tL = 100 nsec, µa ≈ 500 cm-1), the plasma formation threshold is 12-18 J/cm2.56,57 However, for ultrashort pulse durations of ≈100 fsec, linear absorption of the target is nearly irrelevant, even for linear absorption coefficients of as high as µa ≈ 1000 cm-1.53 Small impurities, as are found in tap water, are irrelevant for pulse durations in the psec range and shorter.41,46 This is because the irradiance necessary to complete the ionization avalanche during the short laser pulses is so high that the initial electrons are readily created by multiphoton ionization, and linear absorption does not result in a lowering of the threshold.

The laser pulse duration does not only affect the threshold irradiance, but also the entire dynamics of plasma formation and the irradiance dependence of the free-electron density, as shown in Figures 3 and 4. With nsec pulses in the IR (Fig. 3), no free electrons are formed by impact ionization for irradiance values below the breakdown threshold because no seed electrons created by multiphoton ionization are available. Once the irradiance is sufficiently high to provide a seed electron, the ionization cascade proceeds very rapidly, due to the high irradiance. The electron density increases by nine

orders of magnitude within a small fraction of the laser pulse duration, and actually overshoots the critical electron density of ρcr = 1020 cm-3. This results in an extremely sharp breakdown threshold because either a highly ionized plasma is produced, or no plasma at all. It is important to note that this ‘sharpness’ does not exclude the possibility of pulse-to- pulse variations of the threshold irradiance. These variations are due to the probabilistic nature of the multiphoton-induced generation of seed electrons. With fsec pulses (Fig. 4), there is no lack of multi- photon-induced seed electrons for avalanche ionization, and the onset of plasma formation is, therefore, deterministic. An avalanche is initiated at irradiance values considerably lower than the breakdown threshold, and the free electron density varies continuously with irradiance. Therefore, it is possible to generate any desired free-electron density by an appropriate irradiance.47

It is interesting to note that, even for fsec plasmas, avalanche ionization is the mechanism that produces the majority of the free electrons during the laser pulse (Fig. 4). Multiphoton ionization dominates during the initial part of the pulse but avalanche ionization takes over later, as its rate depends on both the irradiance and free-electron density, whereas the multiphoton ionization rate only depends on irradiance (Equation 1).

Plasma formation above the breakdown threshold

At the breakdown threshold, plasma formation is restricted to the focal region of the laser beam. In contrast, when the laser beam provides an energy in excess of the breakdown threshold and is focused within a transparent medium, the plasma formation is characterized by a growth of the plasma from the beam waist towards the incoming laser beam, as illustrated in Figure 5. Almost no plasma develops behind the laser focus since most of the laser light has already been absorbed prior to and in the beam waist. Thus, the region behind the focus is ‘shielded’ by the plasma absorption.34,48,58-60

A realistic explanation for the expansion of the plasma is provided by the ‘moving breakdown’ model originally proposed by Raizer,61 and further refined by Docchio and coworkers.58,59 In this model, it is assumed that optical breakdown is independent of the preceding plasma formation, and occurs at all locations where the irradiation exceeds the breakdown threshold. As the power increases during the laser pulse, the plasma front moves along the optical axis at the same velocity as the location where the breakdown threshold is exceeded. This process is illustrated in Figure 6. For a Gaussian beam, Docchio and coworkers derived the following prediction for the plasma length zmax from the beam waist towards the incoming laser beam that is reached at the intensity peak of the laser pulse.58

Here zR is the Rayleigh range, and β the ratio between peak irradiance in the laser pulse and the threshold irradiance for laser-induced breakdown, i.e. β = I/Ith. The laser radiation incident on the plasma after the intensity peak of the laser pulse only serves to heat the plasma, but does not elongate it further.

The validity of the moving breakdown model was shown experimentally for Nd:YAG laser pulses (λ = 1064 nm) with pulse durations in the nsec and psec range.48,58 The quantitative predictions were very good for psec pulses, but degrade for pulse durations in the nsec range as the assumption of a spatially and temporally constant breakdown threshold during the laser pulse is incorrect in these conditions.34,48 The reason for this is that the UV emission of the plasma contributes to the formation of additional free electrons in the plasma vicinity that act as seed electrons for cascade ionization. For nsec pulses, this leads to a lowering of the breakdown threshold during the laser pulse and, thus, to a much larger plasma size than predicted by the moving breakdown model. At the start of plasma formation, the threshold is determined by the high irradiance needed to generate the initial quasi-free electrons for the ionization avalanche through multiphoton ionization. Later, when seed electrons are provided by the UV plasma luminescence, the breakdown threshold decreases to the irradiance value necessary to reach the critical electron density by avalanche ionization. For shorter pulses in the psec or fsec range, a higher irradiance is required to achieve the critical free-electron density at the end of the ionization avalanche, and the creation of seed electrons for the avalanche through multiphoton ionization does not provide an additional barrier. Therefore, at superthreshold energy, the breakdown threshold remains constant throughout the laser pulse and the predictions of the moving breakdown model hold.

For breakdown produced using fsec pulses focused at moderate angles, the plasma length observed above the breakdown threshold is considerably longer than the spatial length of the laser pulse (30 µm for a 100fsec pulse). This indicates that plasma formation starts before the pulse reaches the laser focus. The plasma front moves with the laser pulse toward the focus, so that free electrons remain in its wake.62 This is in contrast to psec and nsec breakdown, where the physical length of the pulse is much longer than the plasma and the plasma front moves from the focus

towards the incoming laser beam. The plasma length for fsec pulses is proportional to cvvvvvvβ – 1, as for psec

pulses, even though the plasma front moves in opposite direction.34 In both cases, the extent of the plasma on the laser side is determined by the maximum axial distance from the beam waist at which the breakdown threshold is exceeded.

The situation is completely different for plasmas

formed at free surfaces in air. During the initial phase of the superthreshold laser pulse, plasma formation is restricted to the target material because the breakdown threshold is higher in air than in the tissue.34 When the electron density becomes high enough that the plasma frequency exceeds the frequency of the light (the critical value is 1021 cm-3 for λ = 1064 nm34), the plasma absorption coefficient increases drastically and the plasma becomes highly reflective up to a value of R = 0.9 for very high superthreshold exposures. This change has two consequences:

(a) the plasma electron and energy densities increase very rapidly,66 and (b) plasma formation extends into the surrounding air because hot electrons ejected from the target start to ionize the air. The latter process leads to the development of a plasma plume that largely reduces the amount of laser light reaching the target.35 It is only for fsec laser pulses that the coupling of optical energy into the target is not impaired by plasma shielding, since the laser pulse is too short to allow the formation of a plasma plume during the laser pulse.66

Plasma absorption

Plasma absorption determines how much energy is coupled into the target medium and how much is transmitted past the target volume. As well as influencing the efficiency of the laser surgical process, plasma absorption is important for its safety if surgery is performed near sensitive, strongly absorbing biological structures as, for example, the retina. Absorption coefficients of plasmas produced in bulk water have been determined experimentally by measuring the plasma transmission, scattering and reflection, together with the plasma length.34,58 The investigations covered a range of radiant exposures up to 50 times threshold, and yielded values of between 100 and about 400 cm-1, depending on pulse duration (6 nsec and 30 psec), wavelength, and radiant exposure. Slightly higher values of between 100 and about 1000 cm-1 were obtained by numerical calculations of the plasma absorption coefficients at the breakdown threshold that considered the time evolution of the free-electron density and the absorption cross section of free electrons for IBA. These calculations were performed for pulse durations of between 100 fsec and 100 nsec.46

For plasma formation at tissue surfaces in air, no experimental data for plasma absorption coefficients and the spatial distribution of energy deposition are available to date. Once the electron density at the surface starts to exceed the critical value of ≈1021 cm-3, the absorption coefficients will certainly be higher than the values for bulk media. Feit et al.66 assumed that plasma is formed in a layer with a thickness of only a few nanometers, but did not discuss how the plasma electron and energy densities corresponding to such a small absorption depth relate to the ablation thresholds of soft and hard tissues, respectively.

Plasma energy density

Plasma energy density is closely linked to the strength of the mechanical effects (shock waves and cavitation) associated with breakdown. It determines how strongly disruptive the breakdown event is, and how much mechanical damage is caused in the vicinity of the laser focus. The deposition of optical energy into the medium is mediated by the generation and subsequent acceleration of free electrons. The energy gained by the electrons is transferred to heavy plasma particles through collisions and recombination, resulting in a heating of the atomic and ionic plasma constituents. The number of collisions and recombination events and the resulting energy transfer to the medium are proportional to the laser pulse duration. Therefore, the plasma energy density must increase with increasing laser pulse duration. Theoretical predictions of the dependence of plasma energy density on laser pulse duration using Equation

(1) are shown in Figure 7.46 For fsec exposures, the laser pulse duration is shorter than the electron cooling and recombination times. Thus, minimal energy is transferred during the pulse, and the energy density deposited into the breakdown region is simply given by the number of free electrons produced, multiplied by the mean energy gain of each electron. For pulse durations longer than the electron cooling time (several psec) and recombination time (several 10 psec), a dynamic equilibrium is established between the energy transfer through collision and recombination and the generation of free electrons by the incident radiation. For pulse durations in the nsec range, the calculated energy density is proportional to the laser pulse duration.

Experimental values of the plasma energy density are 33-40 kJ/cm3 for 6-nsec pulses, ≈10 kJ/cm3 for 30-psec pulses, and less than 1 kJ/cm3 for 100fsec pulses.68,69 The model predictions in Figure 7

Fig. 7. Calculated plasma energy density at the optical breakdown threshold versus laser pulse duration. The calculations were performed for a wavelength of 580 nm and a critical electron density of 1020 cm-3, which is realistic for nsec-optical breakdown (see text). For psec and fsec pulses, the critical electron density is approximately 1021 cm-3, and thus the actual plasma energy density is approximately ten times higher than plotted. (Reproduced from Noack and Vogel46 by courtesy of the publisher.)

agree qualitatively well with the experimental data. However, the calculated value for the 6-nsec pulse duration (150 kJ/cm3) is approximately four times greater than the experimental value because the plasma expansion during the laser pulse is not accounted for in the model. On the other hand, the energy density values predicted for 100-fsec and 30-psec pulses (150 J/cm3 and 550 J/cm3, respectively) are by about one order of magnitude smaller than the experimental values. This is due to the fact that all energy density values were calculated assuming an electron density of 1020 cm-3. A much better agreement with the experimental data is obtained assuming an electron density of 1021 cm-3 for psec and fsec breakdown, which also yields the best match between calculated and experimental values for the breakdown thresholds.46

The above data indicate that, while the threshold energy density for optical breakdown and plasmamediated ablation using nsec pulses is extremely high, for fsec pulses it is smaller than the vaporization enthalpy of water at constant pressure, and resembles the volumetric energy density threshold for ablation based on linear absorption.

For plasma-mediated ablation in bulk media at superthreshold energies, the plasma volume grows during the laser pulse, and thus there is little variation in the plasma energy density at superthreshold energies. Details of the dependence of plasma energy density on the laser pulse energy, pulse duration and focusing angle have been analyzed by Vogel et al.34,38,70 The situation differs for plasmas produced with ultrashort laser pulses at surfaces in air. In that case, there is no shielding of the irradiance at the target surface until the breakdown threshold in air has been reached. Thus, the energy density in the superficial plasma layer will increase with growing radiant exposure, due to the increase in IBA events with growing electron density.

Thermo-mechanical and chemical plasma effects

During laser-induced plasma formation, an extraordinarily high energy density develops in the focal volume within a very short time, in particular for nsec and psec optical breakdown.22,48,68,70-74 Temperature and pressure rise rapidly to very high values, causing an explosive expansion of the laser plasma. This expansion of the plasma leads to the production of a shock wave68,75 and, if the application site is in a fluid environment, to the formation of a cavitation bubble,68,75 as shown in Figures 8 and 9. The high initial plasma pressure results in a very rapid bubble expansion that overshoots the equilibrium state where the internal bubble pressure equals the hydrostatic pressure. When both pressures are equal, the kinetic energy of the fluid proximal to the bubble has reached its maximum and, owing to inertia, the bubble continues to expand radially. The expansion leads to a drop in the internal bubble pressure, and

the increasing difference between the hydrostatic pressure and the internal bubble pressure decelerates the expansion and brings it to a halt. At this point, all kinetic energy is transformed into the potential energy of the expanded bubble (Fig. 9b). The bubble energy is related to the radius of the bubble at its maximum expansion, Rmax, and the difference between the hydrostatic pressure p0 and the vapor pressure pv inside the bubble by:75

The expanded bubble collapses again, due to the static background fluid pressure. This collapse compresses the bubble content into a very small volume, thus generating a very high pressure that can exceed 1 Gpa.76 The rebound of the compressed bubble interior leads to the emission of a strong stress transient into the surrounding liquid that for approximately spherical bubbles evolves into a shock wave.22,76

While the events during bubble generation are strongly influenced by the laser parameters, the subsequent bubble dynamics are primarily influenced by the boundary conditions in the neighborhood of the laser focus. In a free fluid, a spherical bubble retains its spherical shape while oscillating, and the bubble collapse takes place at the site of bubble formation. Near material boundaries, the collapse is asymmetric and associated with the formation of one or two high-speed water jets, which concentrate the bubble energy at some distance from the locus of bubble generation.76-79 When the bubble collapses in the vicinity of a rigid boundary (such as, for example, an intraocular lens implant), the jet is directed towards this boundary, as illustrated in Figure 9c. During bubble collapse near elastic tissue-like boundaries, bubble splitting and formation of two liquid jets directed away from and towards the boundary have been observed, and velocities of as high as 960 m/sec have been measured for the jet directed towards the boundary.78

A summary of peak shock pressure amplitudes and transduction efficiencies of absorbed optical energy into cavitation bubble energy for different pulse durations is given in Table 1. Vogel et al. presented a complete energy balance for plasma formation in bulk water at nsec to fsec time scales.34,70 They found that the transduction of laser energy into mechanical energy (shock wave and cavitation bubble energy) for nsec pulses is as high as 90%, more than for any other laser-tissue interaction. The ratio of shock wave energy (not included in Table 1) to cavitation bubble energy was ≈2:1 for nsec pulses and ≈3:2 for psec pulses.

The explosive expansion of the plasma produces disruptive tissue effects extending far beyond the vaporization and disintegration of tissue that occurs within the plasma volume. In the immediate vicinity of the plasma, the effects of the shock wave and

cavitation bubble expansion can hardly be distinguished, but at a somewhat larger distance, cavitation effects are without doubt responsible for the creation of morphologically-identifiable tissue alterations.

The effects of the plasma and bubble expansion strongly depend on the location of the plasma in the tissue. When the plasma is formed in the bulk of the tissue, all deposited energy above the vaporization threshold acts to deform the surrounding tissue.71,84 However, when the laser pulse is focused on a tissue surface in a liquid environment, a large fraction of the deposited energy is imparted to the surrounding fluid, and the hole created in the tissue is only slightly greater than the diameter of the laser focus.34,71 Nevertheless, the inertial confinement of the vaporized material by the surrounding fluid causes a distinct indentation of the tissue surface during the expansion of the cavitation bubble. Collateral effects are much less severe when the plasma is produced at a tissue surface in air where the plasma expansion is not mechanically confined.84,85

As well as the cavitation bubble expansion, the jet formation during bubble collapse is another potent cause of far-reaching tissue effects because the jet concentrates energy at locations away from the site of plasma formation. The jets have been shown to cause collateral damage in photodisruption71 and pulsed laser ablation,79 and to increase the amount of material removed.78,79,86

The reduction of plasma energy density with decreasing pulse duration explains the strong reduction of mechanical effects produced with ultrashort, as opposed to nsec, laser pulses (Table 1). The ratio of mechanical energy Emech to the energy fraction consumed for vaporization of the fluid within

the plasma volume Evap can serve as a metric for the strength of the disruptive effects accompanying

plasma-mediated ablation. The ratio (Emech/Evap) for breakdown in water was found to decrease from 12:1 to 1:2 when the pulse duration was reduced from 6 nsec to 100 fsec.70 The drop in plasma energy density with pulse duration also explains the decrease observed in plasma luminescence, which is no longer visible for pulse durations ≤ 3 psec.62,69

The pressure of the shock waves emitted from the optical breakdown site decreases with decreasing

pulse duration34,68,69,72,73 (Table 1), but not as strongly as the reduction in plasma energy density. The reason for this phenomenon is that psec and fsec plasmas are always produced under stress confinement. While the pressure induced by a phase transition in the plasma volume is small for low energy densities, the thermoelastic stresses are still very high. Numerical simulations for plasmas with a free-electron density of ρ = 1021 cm-3 predicted a temperature rise of 274°C, which is accompanied by the generation of a thermoelastic stress wave with 2.4 kbar compressive amplitude outside the laser focus,47 which is much higher than the saturated vapor pressure for this temperature. The tensile component of the thermoelastic stress wave determines the threshold for bubble formation at the laser focus. We can conclude that the mechanical effects at the threshold for plasma-mediated soft-tissue ablation with fsec pulses are comparable with those accompanying stressconfined linear absorption.

Although high temperatures of several thousand Kelvin are reached within nsec and psec plasmas,87,88 a thermally-modified zone less than 0.2 µm thick was found at the rim of ablation craters in corneal tissue, regardless of pulse duration.71,89 The sharp delineation of the heated zone is due to the sharpness of the plasma boundary, owing to the nonlinearity of its formation and to the short time available for heat diffusion out of the heated volume that is limited by rapid adiabatic cooling during the cavitation bubble expansion.71

Within the plasma, thermal degradation of the tissue is accompanied by chemical dissociation induced by the interaction of the free electrons with the biomolecules and water. Electrons with energies below 15 eV can initiate fragmentation of biomolecules via attachment of the incident electron. The electron attachment leads to the formation of a resonance, namely a transient molecular anion state.47,90,91 For a molecule XY, this process corresponds to XY + e- → XY*-, where the XY*- has a repulsive potential along the X-Y bond coordinate. The molecular anionic state can decay by electron detachment (leaving a vibrationally excited molecule) or by molecular dissociation along one or several specific bonds, such as XY*- → X• + Y-.

Nikogosyan et al.92 describe the formation of OH*

and H2O2 through various pathways following ionization and dissociation of water molecules. Both oxygen species are highly reactive and known to cause cell damage.93 The dissociation of water molecules and free radical formation during nsec and fsec laser-induced plasma formation has been confirmed in several experimental studies.94,95

While the chemical processes within the breakdown region have little practical relevance for laser parameters producing a high plasma energy density, where the tissue effects are dominated by the thermomechanical effects, they are of major importance for fsec plasmas with energy densities below the threshold for bubble formation. Such low-density plasmas are possible because of the gradual increase of the free-electron density with irradiance at ultrashort pulse durations (Fig. 4). They enable highly-local- ized, chemically-mediated ablation or dissection processes with little or no thermomechanical contribution.47

Implications for tissue ablation and intraocular photodisruption

The use of plasma-mediated ablation and disruption that are based on nonlinear absorption enables the performance of surgery inside of transparent biological structures.7,8,27-31,34

Owing to the high energy density at nsec and psec durations, and the inertial confinement inside transparent structures, the precision of plasma-mediated effects is generally compromised by cavitation effects. Therefore, plasma-mediated effects in bulk tissue are better-suited for cutting and disruption than for ablation of large tissue volumes with sharply delineated boundaries.81 The precision of plasmamediated cutting can be optimized by various measures: 1. The use of a clean beam profile that provides the minimum possible spot size at any given focusing angle. 2. The use of the largest possible focusing angle for each application because a large focusing angle guarantees a small focal spot both in lateral and axial direction. 3. Minimization of the aberrations in the optical delivery system, including the contact lenses used for intraocular applications.67,96 4. The use of ultrashort pulse durations to exploit the decrease of the optical breakdown threshold and the minimization of mechanical side-effects observed with decreasing pulse duration. 5. When pulse series are applied, it must be prevented that plasma production is hindered by the cavitation pulses from previous pulses. Hindrance of the plasma production arises when the time between pulses is shorter than the bubble lifetime, for example, when highrepetition rate bursts are used for iridotomies,97 or in intrastromal corneal surgery for refractive corrections. In the latter case, this can be partially circumvented by the use of irradiation strategies that avoid fusion of the microbubbles produced by individual laser pulses to large bubbles that affect the beam path of subsequent laser pulses.29

a

b

c

Fig. 10. Refinement of plasma-mediated tissue effects by the use of shorter laser pulses. a. Cut in Descemet’s membrane in the cornea produced by 6-nsec laser pulses of 1064 nm wavelength. b. Cut produced by 30-psec laser pulses. The scale bars represent a length of 100 µm. (Reproduced from Vogel et al.36 by courtesy of the publisher.) c. Lenticule dissected out of the corneal stroma using 110-fsec pulses of 780 nm wavelength. (Reproduced from Lubatschowski et al.29 by courtesy of the publisher.)

The refinement of tissue effects with decreasing pulse duration has been demonstrated in various experimental studies,28,29,36,84,98,99 and some clinical studies,82,100 and is illustrated in Figure 10. When ultrashort laser pulses are applied at a very large numerical aperture, it is possible to perform intranuclear chromosome dissection in living cells.30 In this procedure, the material removal per pulse is less than 0.1 µm3. On the other hand, standard ophthalmic applications, such as iridotomies and capsulotomies, often take advantage of the disruptive mechanical effects accompanying optical breakdown, and are thus mostly performed with nsec pulses.7

Plasma-mediated processes have also been applied

for material removal from soft and hard tissue surfaces in air. We showed above that the plasma absorption coefficients in bulk aqueous media are two orders of magnitude smaller than the linear tissue absorption coefficients for ArF excimer or Er:YAG laser radiation that are often used for tissue ablation at surfaces. Therefore, the desired precision of plasma-mediated ablation must be controlled by an appropriate choice of focusing angle and pulse energy. Both parameters determine the growth of the plasma from the beam waist into the cone angle of the incoming laser beam. The minimal ablation depth for fsec ablation of soft tissues was found to be relatively large (1.5-200 µm) when the laser radiation was focused to a spot size on the order of 20 µm. While the ablation depth could probably be reduced by using a smaller spot size (i.e., a larger focusing angle) and very small pulse energies, this would involve a considerably longer processing time. Moreover, such a procedure would require very precise tracking of the tissue surface in the axial direction of the laser beam, because the focal spot must be located exactly at the tissue surface in order to achieve precise ablation.

When the laser beam is only moderately focused, a tissue layer of up to 200 µm in thickness is ejected by the expansion of the plasma formed in the beam waist.74,104,105 Because mechanical ejection is involved in the ablation process, the ablation depth depends on the ultimate tensile strength of the tissue. At equal radiant exposure and pulse duration, much smaller ablation rates have been observed for the mechanically resistant corneal stroma (1.5-10 µm with 30-psec pulses89 and 3-16 µm with 140-fsec pulses103) than for the much weaker neural tissue (50200 µm with 30-psec pulses105, 20-200 µm with 140fsec pulses104).

The ablation efficiency achieved using fsec pulses is larger than when using psec and nsec pulses.103,104 This trend can be explained by the decrease of plasma energy density with shorter pulse durations.

The nonlinear dependence of the plasma energy density with irradiance (Figs. 3 and 4) enables a large ablation depth to be achieved with little thermal damage to the residual tissue. Thermal side-effects are almost negligible with single laser exposures, regardless of the laser pulse duration.71,89,104 Even with nsec pulses, the thickness of the thermallydamaged layer in corneal tissue remains < 1 µm.71 However, when ultrashort laser pulses are applied at high repetition rates, the residual heat remaining in the nonablated tissue may accumulate and lead to larger zones of thermal damage if the laser beam is applied to a large spot and not laterally scanned during ablation.66,102,106 This phenomenon is due to the continuous dependence of free-electron density and energy density on irradiance in the fsec range (Fig. 4). Therefore, when using fsec pulses, the target is heated in regions where the ablation threshold is not exceeded. When nsec pulses are used, the sharp drop of electron density below the optical breakdown

threshold prevents heating of the non-ablated tissue (Fig. 3).

The dependence of the ablation depth on radiant exposure was found to be linear for neural tissue,104 but logarithmic for corneal tissue.89 The ablation efficiency for corneal tissue is thus highest close to the ablation threshold, and decreases for higher radiant exposures.53,103 This dependence is typical of a blow-off ablation process under conditions where the spatial energy density profile in the tissue corresponds to Beer’s law (exponential decrease of the energy density with depth). Similar conditions may arise in plasma-mediated ablation due to the shielding of deeper tissue layers by the plasma produced at the surface. The shielding-induced reduction in energy density with increasing depth is, to a certain degree, counteracted by self-focusing and self-chan- nelling effects, leading to filament formation beyond the location of the laser focus. This phenomenon has been described theoretically first for laser beam propagation in air,107,108 but was also observed in water,70 gelatin,109 and in TEM pictures of fsec la- ser-irradiated cornea.29 Filament formation could possibly explain why the ablation depth for neural tissue is much larger than for corneal tissue, and exhibits a different dependence on radiant exposure. The thin filament, much smaller than the irradiated spot size, may only be able to remove mechanicallyweak material like neural tissue, but hardly influences strong materials like the corneal stroma.

Conclusions

The most outstanding feature of plasma-mediated ablation is its ability to create ablation effects inside tissues that are transparent at low irradiance. On the other hand, plasma-mediated surface ablation of soft tissues enables the combination of relatively large etch depths with minimal thermal damage, and thus provides an additional unique feature. For both applications, minimization of mechanical collateral effects can be achieved by employing ultrashort laser pulses. However, if disruptive laser effects are desired, as, for example, in posterior capsulotomies, the use of nsec pulses is more appropriate.

References

1.Krasnov MM: Q-switched laser iridectomy and Q-switched laser goniopuncture. Adv Ophthalmol 34:192-196, 1977

2.Aron-Rosa D, Aron J, Griesemann J, Thyzel R: Use of the neodymium:YAG laser to open the posterior capsule after lens implant surgery. J Am Interocul Implant Soc 6:352354, 1980

3.Van der Zypen E, Fankhauser F, Bebie H, Marshall J: Changes in the ultrastructure of the iris after irradiation with intense light. Adv Ophthalmol 39:59-180, 1979

4.Fankhauser F, Roussel P, Steffen J, Van der Zypen E, Chrenkova A: Clinical studies on the efficiency of high

power laser radiation of some structures of the anterior segment of the eye. Int Ophthalmol 3:129-139, 1981

5.Fankhauser F, Loertscher HP, Van der Zypen E: Clinical studies on high and low laser power radiation upon some structures of the anterior and posterior segments of the eye. Int Ophthalmol 5:15-32, 1982

6.Trokel SL (ed): YAG Laser Ophthalmic Microsurgery. Norwalk, CN: Appleton-Century-Crofts 1983

7.Steinert RF, Puliafito CA: The Nd:YAG Laser in Ophthalmology: Principles and Clinical Practice of Photodisruption. Philadelphia, PA: Saunders 1986

8.Fankhauser F, Kwasniewska S: Neodymium:yttrium-alu- minum-garnet laser. In: L’Esperance FA (ed) Ophthalmic Lasers, 3rd edn, pp 781-886. St Louis, MO: CV Mosby 1989

9.Gabel VP: Lasers in vitreoretinal therapy. Ann Ophthalmic Laser Surg 1:75-89, 1992

10.Mainster MA, Sliney DH, Belcher CD, Buzney SM: Laser photodisruptors: damage mechanisms, instrument design and safety. Ophthalmology 90:973-991, 1983

11.Gabel VP, Neubauer L, Zink H, Birngruber R: Ocular side effects following neodymium:YAG laser irradiation. Int Ophthalmol Clin 25:137-149, 1985

12.Taboada J: Interaction of short laser pulses with ocular tissue. In: Trokel SL (ed) YAG Laser Ophthalmic Microsurgery, pp 15-38. Norwalk, CN: Appleton-Century-Crofts 1983

13.Puliafito CA, Steinert RF: Short-pulsed Nd:YAG laser microsurgery of the eye: biophysical considerations. IEEE J Quant Electr QE-20:1442-1448, 1984

14.Loertscher HP: Laser-induced breakdown for ophthalmic applications. In: Trokel SL (ed) YAG Laser Ophthalmic Microsurgery, pp 39-66. Norwalk, CN: Appleton-Century- Crofts 1983

15.Docchio F, Dossi L, Sacchi CA: Q-switched Nd:YAG laser irradiation of the eye and related phenomena: an experimental study I. Optical breakdown determination for liquids and membranes. Lasers Life Sci 1:87-103, 1986

16.Docchio F, Sachhi CA, Marshall J: Experimental investigation of optical breakdown thresholds in ocular media under single pulse irradiation with different pulse durations. Lasers Ophthalmol 1:83-93, 1986

17.Steinert RF, Puliafito CA: Plasma formation and shielding by three ophthalmic neodymium-YAG lasers. Am J Ophthalmol 96:427-434, 1983

18.Steinert RF, Puliafito CA Kittrell C: Plasma shielding by Q-switched and mode-locked Nd:YAG lasers. Ophthalmology 90:1003-1006, 1983

19.Capon MRC; Docchio F, Mellerio J: Nd:YAG laser photodisruption: an experimental investigation on shielding and multiple plasma formation. Graefe’s Arch Clin Exp Ophthalmol 226:362-366, 1988

20.Docchio F, Sacchi CA: Shielding properties of laser-in- duced plasmas in ocular media irradiated by single Nd:YAG pulses of different durations. Invest Ophthalmol Vis Sci 29:437-443, 1988

21.Fujimoto JG, Lin WZ, Ippen EP, Puliafito CA, Steinert RF: Time-resolved studies of Nd:YAG laser-induced breakdown. Invest Ophthalmol Vis Sci 26:1771-1777, 1985

22.Vogel A, Hentschel W, Holzfuss J, Lauterborn W: Cavitation bubble dynamics and acoustic transient generation in ocular surgery with pulsed neodymium:YAG lasers. Ophthalmology 93:1259-1269, 1986

23.Capon MRC, Mellerio J: Nd:YAG Lasers: plasma characteristics and damage mechanisms. Lasers Ophthalmol 1:95106, 1986

24.Mellerio J, Capon MRC, Docchio F: Nd:YAG lasers: a potential hazard from cavitation bubble behaviour in an-

terior chamber procedures? Lasers Ophthalmol 1:185-190, 1987

25.Zysset B, Fujimoto JG, Deutsch TF: Time-resolved measurements of picosecond optical breakdown. Appl Phys B 48:139-147, 1989

26.Doukas AG, Zweig AD, Frisoli JK, Birngruber R, Deutsch T: Non-invasive determination of shock wave pressure generated by optical breakdown. Appl Phys B 53:237-245, 1991

27.Niemz MH, Hoppeler TP, Juhasz T, Bille JF: Intrastromal ablations for refractive corneal surgery using picosecond infrared laser pulses. Lasers Light Ophthalmol 5:149-155, 1993

28.Juhasz T, Loesel FH, Kurtz RM, Horvath C, Bille JF, Mourou G: Corneal refractive surgery with femtosecond lasers. IEEE J Sel Top Quant Electr 5:902-910, 1999

29.Lubatschowski H, Heisterkamp A, Will F, Serbin J, Bauer T, Fallnich C, Welling H: Ultrafast laser pulses for medical applications. Proc SPIE 4633:38-49, 2002

30.König K, Riemann I, Fischer P, Halbhuber K: Intracellular nanosurgery with near infrared femtosecond laser pulses. Cell Mol Biol 45:195-201, 1999

31.Venugopalan V, Guerra A, Nahen K, Vogel A: Role of laser-induced plasma formation in pulsed cellular microsurgery and micromanipulation. Phys Rev Lett 88:078103, 2002

32.Shen YR: The Principles of Nonlinear Optics. New York, NY: Wiley 1984

33.Vogel A: Nonlinear absorption: intraocular microsurgery and lithotripsy. Phys Med Biol 42:895-912, 1997

34.Vogel A: Optical Breakdown in Water and Ocular Media, and its Use for Intraocular Photodisruption. Aachen: Shaker 2001

35.Root RG: Modeling of post-breakdown phenomena. In: Radziemski LJ, Cremers DA (eds) Laser-Induced Plasmas and Applications, pp 69-103. New York, NY: Marcel Dekker 1989

36.Vogel A, Capon MRC, Asiyo MN, Birngruber R: Intraocular photodisruption with picosecond and nanosecond laser pulses: tissue effects in cornea, lens and retina. Invest Ophthalmol Vis Sci 35:3033-3044, 1994

37.Sacchi CA: Laser-induced electric breakdown in water. J Opt Soc Am B 8:337-345, 1991

38.Williams F, Varama SP, Hillenius S: Liquid water as a lonepair amorphous semiconductor. J Chem Phys 64:1549-1554, 1976

39.Ready JF: Effects of High-Power Laser Radiation. Orlando, FL: Academic Press 1971

40.Bloembergen N: Laser-induced electric breakdown in solids. IEEE J Quant Electr QE-10:375-386, 1974

41.Kennedy PK: A first-order model for computation of la- ser-induced breakdown thresholds in ocular and aqueous media. Part 1. Theory. IEEE J Quant Electr QE-31:2241- 2249, 1995

42.Niemz MH: Threshold dependence of laser-induced optical breakdown on pulse duration. Appl Phys Lett 66:11811183, 1995

43.Stuart BC, Feit MD, Hermann S, Rubenchik AM, Shore BW, Perry MD: Nanosecond to femtosecond laser-induced breakdown in dielectrics. Phys Rev B 53:1749-1761, 1996

44.Feng Q, Moloney JV, Newell AC, Wright EM, Cook K, Kennedy PK, Hammer DX, Rockwell BA, Thompson CR: Theory and simulation on the threshold of water breakdown induced by ultrashort laser pulses. IEEE J Quant Electr 33:127-137, 1997

45.Lenzner M, Krüger J, Sartania S, Cheng Z, Spielmann C, Mourou G, Kautek W, Krausz F: Femtosecond optical breakdown in dielectrica. Phys Rev Lett 80:4076-4079, 1998

46.Noack J, Vogel A: Laser-induced plasma formation in water at nanosecond to femtosecond time scales: calculation of thresholds, absorption coefficients, and energy density. IEEE J Quant Electr 35:1156-1167, 1999

47.Vogel A, Noack J, Hüttmann G, Paltauf G: Femtosecond- laser-produced low-density plasmas in transparent biological media: a tool for the creation of chemical, thermal and thermomechanical effects below the optical breakdown threshold. Proc SPIE 4633:23-37, 2002

48.Vogel A, Nahen K, Theisen D: Plasma formation in water by picosecond and nanosecond Nd:YAG laser pulses. Part 1. Optical breakdown at threshold and superthreshold irradiance. IEEE J Sel Top Quant Electr 2:847-860, 1996

49.Hammer DX, Noojin GD, Thomas RJ, Hopkins RA, Kennedy PK, Druessel JJ, Rockwell BA, Welch AJ, Cain CP: Measurement of the self-focusing threshold in aqueous media by ultrashort laser pulses. SPIE Proc 2975:163-172, 1997

50.Schaffer CB, Brodeur A, García JF, Mazur E: Micromachining bulk glass by use of femtosecond laser pulses with nanojoule energy. Opt Lett 26:93-95, 2001

51.Barnes PA, Rieckhoff KE: Laser induced underwater sparks. Appl Phys Lett 13:282-284, 1968

52.Hammer DX, Thomas RJ, Noojin GD, Rockwell BA, Kennedy PA, Roach WP: Experimental investigation of ultrashort pulse laser-induced breakdown thresholds in aqueous media. IEEE J Quant Electr QE-3:670-678, 1996

53.Oraevsky AA, DaSilva LB, Rubenchik AM, Feit MD, Glinsky ME, Perry MD, Mammini BM, Small W IV, Stuart BC: Plasma mediated ablation of biological tissues with nanosecond to femtosecond laser pulses: relative role of linear and nonlinear absorption. IEEE J Sel Top Quant Electr 2:801-809, 1996

54.Pettit GH, Ediger MN: Corneal tissue absorption coefficients for 193and 213-nm ultraviolet radiation. Appl Opt 35:3386-3391, 1996

55.Venugopalan V, Nishioka NS, Mikic BB: The thermodynamic response of soft biological tissues to pulsed ultraviolet laser radiation. Biophys J 69:1259-1271, 1995

56.Walsh JT, Deutsch TF: Pulsed CO2 laser tissue ablation: measurement of the ablation rate. Lasers Surg Med 8:264-

275, 1988

57.Green HA, Domankewitz Y, Nishioka N: Pulsed carbon dioxide laser ablation of burned skin: in vitro and in vivo analysis. Lasers Surg Med 10:476-484, 1990

58.Docchio F, Regondi P, Capon MRC, Mellerio J: Study of the temporal and spatial dynamics of plasmas induced in liquids by nanosecond Nd:YAG laser pulses. 1. Analysis of the plasma starting times. Appl Opt 27:3661-3674, 1988

59.Docchio F, Regondi P, Capon MRC, Mellerio J: Study of the temporal and spatial dynamics of plasmas induced in liquids by nanosecond Nd:YAG laser pulses. 2. Plasma luminescence and shielding. Appl Opt 27:3669-3674, 1988

60.Nahen K, Vogel A: Plasma formation in water by picosecond and nanosecond Nd:YAG laser pulses. Part 2. Plasma transmission, scattering and reflection. IEEE J Sel Top Quant Electr 2:861-871,1996

61.Raizer YP: Heating of a gas by a powerful light pulse. Sov Phys JETP 48:1508-1519, 1965

62.Kennedy PK, Hammer DX, Rockwell BA: Laser-induced breakdown in aqueous media. Progr Quant Elect 21:155248, 1997

63.Hughes TP: Plasmas and Laser Light. Bristol: Adam Hilger 1975

64.Godwin RP, Van Kessel CGM, Olsen JN, Sachsenmaier P, Sigel R: Reflection losses from laser-produced plasmas. Z Naturforsch 32a:1100-1107, 1977

65.Rubenchik AM, Feit MD, Perry MD, Larsen JT: Numerical simulation of ultra-short laser pulse energy deposition

and transport for material processing. Appl Surf Sci 127/ 129:193-198, 1998

66.Feit MD, Rubenchik AM, Kim BM, Da Silva LB, Perry MD: Physical characterization of ultrashort laser pulse drilling of biological tissue. Appl Surf Sci 127/129:869874, 1998

67.Vogel A, Nahen K, Theisen D, Birngruber R, Thomas RJ, Rockwell BA: Influence of optical aberrations on laserinduced plasma formation in water and their consequences for intraocular photodisruption. Appl Opt 38:3636-3643, 1999

68.Vogel A, Busch S, Parlitz U: Shock wave emission and cavitation bubble generation by picosecond and nanosecond optical breakdown in water. J Acoust Soc Am 100:148165, 1996

69.Noack J, Hammer DX, Noojin GD, Rockwell BA, Vogel

A:Influence of pulse duration on mechanical effects after laser-induced breakdown in water. J Appl Phys 83:74887495, 1998

70.Vogel A, Noack J, Nahen K, Theisen D, Busch S, Parlitz U, Hammer DX, Noojin GD, Rockwell BA, Birngruber R: Energy balance of optical breakdown in water at nanosecond to femtosecond time scales. Appl Phys B 68:271-280, 1999

71.Vogel A, Schweiger P, Frieser A, Asiyo M, Birngruber

R:Intraocular Nd:YAG laser surgery: light-tissue interaction, damage range, and reduction of collateral effects. IEEE J Quant Electr 26:2240-2260, 1990

72.Juhasz T, Hu XH, Turi L, Bor Z: Dynamics of shock waves and cavitation bubbles generated by picosecond laser pulses in corneal tissue and water. Lasers Surg Med 15:91-98, 1994

73.Juhasz T, Kastis GA, Suarez C, Bor Z, Bron WE: Timeresolved observations of shock waves and cavitation bubbles generated by femtosecond laser pulses in corneal tissue and water. Lasers Surg Med 19:23-31, 1996

74.Fischer JP, Juhasz T, Bille JF: Time resolved imaging of the surface ablation of soft tissue with IR picosecond laser pulses. Appl Phys A 64:181-189, 1997

75.Cole RH: Underwater Explosions. Princeton, NJ: Princeton University Press 1948

76.Vogel A, Lauterborn W, Timm R: Optical and acoustic investigation of the dynamics of laser-produced cavitation bubbles near a solid boundary. J Fluid Mech 206:299-338, 1989

77.Lauterborn W, Bolle H: Experimental investigations of cavitation bubble collapse in the neighborhood of a solid boundary. J Fluid Mech 72:391-399, 1975

78.Brujan EA, Nahen K, Schmidt P, Vogel A: Dynamics of laser-induced cavitation bubbles near an elastic boundary. J Fluid Mech 433:251-281, 2001

79.Brujan EA, Nahen K, Schmidt P, Vogel A: Dynamics of laser-induced cavitation bubbles near elastic boundaries: influence of the elastic modulus. J Fluid Mech 433:283314, 2001

80.Vogel A, Busch S, Jungnickel K, Birngruber R: Mechanisms of intraocular photodisruption with picosecond and nanosecond laser pulses. Lasers Surg Med 15:32-43, 1994

81.Vogel A, Günther T, Asiyo-Vogel M, Birngruber R: Factors determining the refractive effects of intrastromal photorefractive keratectomy with the picosecond laser. J Cataract Refract Surg 23:1301-1310, 1997

82.Juhasz T, Kurtz R, Horvath C, Suarez C, Nordan LT, Slade SG: Femtosecond laser eye surgery: the first clinical experience. Proc SPIE 4633:1-10, 2002

83.Birngruber R, Hillenkamp F, Stefani FH, Gabel VP: Q- switched ruby laser damage of the rabbit eye lens. Adv Ophthalmol 34:158-163, 1977

84.Stern D, Schoenlein RW, Puliafito CA, Dobi ET, Birngruber R, Fujimoto JG: Corneal ablation by nanosecond, picosecond, and femtosecond lasers at 532 and 625 nm. Arch Ophthalmol 107:587-592, 1989

85.Fabbro R, Fournier J, Ballard P, Devaux D, Virmont J: Physical study of laser-produced plasma in confined geometry. J Appl Phys 68:775-784, 1990

86.Chapyak EJ, Godwin RP: Simulations of laser thrombolysis. Proc SPIE 3590:328-335, 1999

87.Stolarski DJ, Hardman J, Bramlette CG, Noojin GD, Thomas RJ, Rockwell BA, Roach WP: Integrated light spectroscopy of laser-induced breakdown in aqueous media. SPIE Proc 2391:100-109, 1995

88.Chapyak EJ, Godwin RP, Vogel A: A comparison of numerical simulations and laboratory studies of shock waves and cavitation bubble growth produced by optical breakdown in water. Proc SPIE 2975:335-342, 1997

89.Niemz MH, Klancnik EG, Bille JF: Plasma-mediated ablation of corneal tissue at 1053 nm using a Nd:YLF oscillator/regenerative amplifier laser. Lasers Surg Med 11: 426-431, 1991

90.Boudaiffa B, Cloutier P, Hunting D, Huels MA, Sanche L: Resonant formation of DNA strand breaks by low-en- ergy (3 to 20 eV) electrons. Science 287:1658-1660, 2000

91.Hotop H: Dynamics of low energy electron collisions with molecules and clusters. In: Christophorou LG, Olthoff JK (eds) Proceedings International Symposium on Gaseous Dielectrics IX, May 22-25, 2001, Ellicott City, MD, pp 3- 14. New York, NY: Kluwer Academic/Plenum Press 2001

92.Nikogosyan DN, Oraevsky AA, Rupasov V: Two-photon ionization and dissociation of liquid water by powerful laser UV radiation. Chem Phys 77:131-143, 1983

93.Tirlapur UK, König K, Peuckert C, Krieg R, Halbhuber KJ: Femtosecond near-infrared laser pulses elicit generation of reactive oxygen species in mammalian cells leading to apoptosis-like death. Exp Cell Res 263:88-97, 2001

94.Timberlake GT, Gemperli AW, Larive CK, Warren KA, Mainster MA: Free radical production by Nd:YAG laser photodisruption. Ophthalmic Surg Lasers 28:582-589, 1997

95.Heisterkamp A, Ripken T, Mamom T, Drommer W, Welling H, Ertmer W, Lubatschowski H: Nonlinear side effects of fs pulses inside corneal tissue during photodisruption. Appl Phys B 74:419-425, 2002

96.Rol P, Fankhauser F, Kwasniewska S: Evaluation of contact lenses for laser therapy. Part 1. Lasers Ophthalmol 1:1- 20, 1986

97.Jungnickel K, Vogel A: Efficiency of bursts in intraocular Nd:YAG laser surgery. Lasers Light Ophthalmol 5:95-99, 1992

98.Zysset B, Fujimoto JG, Puliafito CA, Birngruber R, Deutsch TF: Picosecond optical breakdown: tissue effects and reduction of collateral damage. Lasers Surg Med 9:183-204, 1989

99.Lin CP, Weaver YK, Birngruber R, Fujimoto JG, Puliafito CA: Intraocular microsurgery with a picosecond Nd:YAG laser. Lasers Light Ophthalmol 15:44-53, 1994

100.Geerling G, Roider J, Schmidt-Erfurth U, Nahen K, ElHifnawi ES, Laqua H, Vogel A: Initial clinical experience with the picosecond Nd:YAG laser for intraocular therapeutic applications. Br J Ophthalmol 82:504-509, 1998

101.Niemz MH: Laser-Tissue Interactions: Fundamentals and Applications, 2nd edn. Berlin: Springer 2002

102.Neev J, Da Silva LB, Feit MD, Perry MD, Rubenchik AM, Stuart BC: Ultrashort pulse lasers for hard tissue ablation. IEEE J Sel Top Quant Electr 2:790-800, 1996

103.Kurtz RM, Elner V, Liu X, Juhasz T: Photodisruption in the cornea as a function of pulse width. J Refract Surg 13:653-658, 1997

104.Loesel FH, Fischer JP, Götz MH, Horvath C, Juhasz T, Noack F, Suhm N, Bille JF: Non-thermal ablation of neural tissue with femtosecond laser pulses. Appl Phys B 66:121-128, 1998

105.Fischer JP, Dams J, Götz MH, Kerker E, Loesel FH, Messer CJ; Niemz MH, Suhm N, Bille JF: Plasma-mediated ablation of brain tissue with picosecond laser pulses. Appl Phys B 58:493-499, 1994

106.Kim BM, Feit MD, Rubenchik AM, Joslin EJ, Eichler J, Stoller P, Da Silva LB: Effects on high repetition rate and beam size on hard tissue damage due to subpicosecond laser pulses. Appl Phys Lett 76:4001-4003, 2000

107.Braun A, Korn G, Liu X, Du D, Squier J, Mourou G: Selfchanneling of high-peak-power femtosecond laser pulses in air. Opt Lett 20:73-75, 1995

108.Couairon A, Tzortzakis S, Berge L, Franco M, Prade B, Mysyrowicz A: Infrared femtosecond light filaments in air: simulations and experiments. J Opt Soc Am B 19:11171131, 2002

109.Maatz G, Heisterkamp A, Lubatschowski H, Barcikowski S, Fallnich C, Welling H, Ertmer W: Chemical and physical side effects at application of ultrashort laser pulses for intrastromal refractive surgery. J Opt A: Pure Appl Opt 2:59-64, 2000