Intraocular Photodisruption With Picosecond and Nanosecond Laser

Intraocular Photodisruption With Picosecond and
Nanosecond Laser Pulses: Tissue Effects in
Cornea, Lens, and Retina
Alfred Vogel*\ Malcolm R. C. Capon,% Mary N. Asiyo-Vogel,% and Reginald Birngruber*-\
Purpose. Nd:YAG laser photodisruption with nanosecond (ns) pulses in the millijoule range is
an established tool for intraocular surgery. This study investigates tissue effects in cornea, lens,
and retina to assess whether picosecond (ps) pulses with energies in the microjoule range can
increase the surgical precision, reduce collateral damage, and allow applications requiring
more localized tissue effects than can be achieved with ns pulses.
Methods. Both ps and ns Nd:YAG laser effects on Descemet's membrane, in the corneal
stroma, in the lens, and at the retina were investigated in vitro in bovine and sheep eyes and in
cataractous human lens nuclei. For each tissue, the optical breakdown threshold was determined.
The morphology of the tissue effects and the damage range of the laser pulses were
examined by light and scanning electron microscopy. The cavitation bubble dynamics during
the formation of corneal intrastromal laser effects were documented by time-resolved photography.
Results. The optical breakdown threshold for ps pulses in clear cornea, lens, and vitreous is, on
average, 12 times lower than that for ns pulses. In cataractous lens nuclei, it is lower by a factor
of 7. Using ps pulses, Descemet's membrane could be dissected with fewer disruptive side
effects than with ns pulses, whereby the damage range decreased by a factor of 3. The range
for retinal damage was only 0.5 mm when 200 /xj P s pulses were focused into the vitreous.
Picosecond pulses could be used for corneal intrastromal tissue evaporation without damaging
the corneal epithelium or endothelium, when the pulses were applied in the anterior part of
the stroma. The range for endothelial damage was 150 nm at 80 n] pulse energy. Intrastromal
corneal refractive surgery is compromised by the laser-induced cavitation effects. Tissue displacement
during bubble expansion is more pronounced than tissue evaporation, and irregular
bubble formation creates difficulties in producing predictable refractive changes.
Conclusions. The use of ps pulses improves the precision of intraocular Nd:YAG laser surgery
and diminishes unwanted disruptive side effects, thereby widening the field of potential applications.
Promising fields for further studies are intrastromal corneal refractive surgery, cataract
fragmentation, membrane cutting, and vitreolysis close to the retina. Invest Ophthalmol
Vis Sci. 1994;35:3032-3044.
Oince the early 1980s, photodisruption with Nd:YAG
laser pulses has become a well-established tool for intraocular
surgery. 1 " 4 At present, most clinical photo-
From the *H. Wacker Laboratory for Medical Laser Applications, University Eye
Hospital Munich, Munich; the \Medical Laser Center, Liibeck; %Strathfield, NSW,
Australia; arid the ^University Eye Hospital, Medical University of Liibeck,
Liibeck, Germany.
Supported by German Research Foundation grant no. Bi-321/2-1 (DFG).
Submitted for publication May 4, 1993; revised February 7, 1994; accepted
February 10, 1994.
Proprietary interest category: N.
Reprint requests to: Alfred Vogel, PhD, Medical Laser Center Liibeck, Peter-
Monnik-Weg 4, 23562 Liibeck, Germany.
3032
disruptors deliver laser pulses with a duration of 6 to
12 ns and a typical pulse energy of 1 to 10 mj. With
these instruments, posterior capsulotomy and iridotomy
have become clinical routine, and various other
procedures such as pupillary membranectomy, synechialysis,
and vitreolysis have also been successfully
performed. 34 There are, however, limitations in the
applicability of Nd:YAG laser photodisruption because
its working mechanism is that of a microexplosion
with a large potential for collateral damage in the
surrounding of the application site. The microexplo-
Investigative Ophthalmology & Visual Science, June 1994, Vol. 35, No. 7
Copyright © Association for Research in Vision and Ophthalmology

Tissue Effects of Picosecond and Nanosecond Photodisruption 3033
sion is initiated by the generation of plasma at the laser
focus with a maximum temperature of about 10,000
K. 5 The plasma formation (optical breakdown) is the
primary and intended surgical mechanism because it
leads to evaporation of the material within the focal
region of the laser beam and thus can be used for
tissue cutting. The fast temperature rise during plasma
formation is, however, inevitably accompanied by an
equally fast pressure rise of about 20 kbar, 6 leading to
rapid expansion of the plasma. The plasma expansion
drives a shock wave and expansion of a cavitation bubble,
which causes tissue disruption around the site of
plasma generation. 6 - 7 The disruptive effects may sometimes
contribute to the surgical aim, but they are also
the main source of unwanted side effects of intraocular
Nd:YAG laser surgery. 7 Because the damage range
of the collateral effects depends on the laser pulse energy,
78 a considerable improvement of the precision
of the tissue effects can only be achieved by reducing
the laser pulse energy far below the values currently
applied in clinical practice. This is impossible with
nanosecond pulses requiring a minimal pulse energy
of about 1 mj for plasma formation. To ensure that
plasma formation still occurs with a much smaller
pulse energy, the pulse duration must be reduced to
the picosecond range. 6910
When the pulse duration is reduced from 6 ns to
30 ps, the threshold for plasma formation in distilled
water decreases by a factor of 13 to a value of only 15
ix]. 6 With pulse energies at this low level, the shock
waves have a smaller amplitude and decay much faster
than those generated with nanosecond pulses, and cavitation
is also strongly diminished. 6 It becomes feasible
to perform laser surgery with a repetition rate of 10 to
100 Hz, whereby in each time interval a similar
amount of energy is delivered into the eye, such as
when a few single shots with ns pulses in the mj range
are used. In this way, the amount of tissue evaporated
per unit time (i.e., the cutting efficiency) stays approximately
the same, but the collateral damage is less. Repetitive
ps pulses may, therefore, serve as an intraocular
"laser scalpel" suitable for tissue dissection rather
than tissue disruption. This widens the field of
Nd:YAG laser applications to surgical procedures requiring
more localized tissue effects than those attainable
with ns pulses. Possible examples are vitreous surgery
close to the retina, trabeculopuncture, cataract
fragmentation before surgery, and intrastromal corneal
refractive surgery. 11
This study investigates in vitro the tissue effects of
repetitive ps pulses in cornea, lens, and retina and
compares them to the effects of ns pulses. The aim is to
demonstrate the increased precision in intraocular microsurgery
achievable with ps pulses and to examine
their potential use for some of the abovementioned
new applications. For each tissue investigated, the
threshold for plasma formation was determined with
both pulse durations. To compare the precision of ps
and ns laser surgery, incisions were produced in
Descemet's membrane of bovine eyes, with the corneal
endothelium a sensitive detector of collateral damage.
To analyze the possibility and predictability of intrastromal
corneal refractive surgery, intrastromal laser
effects were generated in bovine and sheep cornea,
and the range for endothelial damage was determined.
With regard to cataract fragmentation, ps and ns effects
in bovine lenses and cataractous human lens nuclei
were compared. For estimating the proximity to
the retina in which vitreous laser surgery can be performed,
we determined the retinal damage range for
ps pulses focused into the vitreous of bovine eyes. The
tissue effects were analyzed by histologic examination
with light microscopy and by scanning electron microscopy.
The histologic findings display the endpoint of the
complex photodisruptive process consisting of plasma
formation, shock wave emission, and cavitation, and
knowledge of that process is necessary to interpret
these findings. Special attention should be paid to the
analysis of the cavitation bubble dynamics because earlier
studies 6712 " 14 indicate that cavitation is the main
source for disruptive effects during intraocular microsurgery.
All investigations reported to date were restricted
to the bubble dynamics in water, so the question
arises whether their results can be transferred to
the bubble dynamics in ocular tissues. The bubble dynamics
in the aqueous fluid of the eye and in the vitreous
is expected to resemble the dynamics in water because
the physical properties of the liquids are similar.
The cavitation bubble dynamics in corneal tissue or
within the lens, however, will certainly be different because
of the fibrous structure and the high viscosity of
these tissues. The bubble dynamics in these tissues are
of interest in relation to intrastromal refractive surgery
of the cornea and to cataract fragmentation. We
have, therefore, studied the cavitation effects occurring
within the corneal stroma by time-resolved photography
to complement our histologic investigations.
MATERIALS AND METHODS
Laser and Application System
The Nd.YAG laser system (YG 671-10, Continuum,
Santa Clara, CA) emits either ns pulses or ps pulses,
both with a repetition rate of 10 Hz and a wavelength
of 1064 nm. The ns pulses have a duration of 6 ns and
an energy of up to 250 mj. The pulse-to-pulse stability
of the energy is better than ±2%. The intensity profile
of the laser beam is nearly gaussian, but it exhibits a
weak ring structure modulating the gaussian profile.

3034 Investigative Ophthalmology 8c Visual Science, June 1994, Vol. 35, No. 7
The ps pulses have a duration of 30 ps and a pulse
energy of up to 10 mj. The pulse-to-pulse fluctuations
of the energy are in the range of ±10%. The beam
profile is gaussian. The pulse energy can be adjusted
continuously for both pulse durations.
The delivery system for the laser pulses is depicted in
Figure 1. The laser beam is expanded, collimated by an
Nd:YAG laser achromat, and coupled into the optics
of a slit lamp microscope by a dichroic mirror. The
mirror is highly reflective for the Nd:YAG laser wavelength
and acts as a beam splitter for visible light. The
front lens of the slit lamp microscope was removed
and replaced by a second Nd:YAG achromat located
behind the mirror-beamsplitter. This achromat takes
up the function of the front lens of the slit lamp microscope
and is at the same time the focusing lens for the
Nd:YAG laser pulses. Aiming is facilitated by a helium-neon
laser beam coupled into the beam path of
the Nd:YAG laser. For investigating the tissue effects,
either enucleated eyes were mounted into an eye
holder or tissue specimens were immersed into a cuvette
(50 X 50 X 50 mm 3 ) filled with saline. To minimize
the spherical aberrations and to simulate the focusing
conditions during clinical Nd:YAG laser applications,
a contact lens (RAK, Rodenstock, Ottobrunn,
Germany) was built into the wall of the cuvette. The
convergence angle in water was 18° for the ps pulses
and 21° for the ns pulses, and the diffraction limited
spot diameter was calculated to be 4.3 /irn for ps pulses
and 3.7 nm for ns pulses. The energy of the laser
pulses was measured with a pyroelectric energy-meter
(Rj 7100, Laser Precision, Utica, NY).
photodiode
energy
meter
aiming laser
Determination of Thresholds for Plasma
Formation
Thresholds for plasma formation with ps and ns pulses
were determined in the cornea, lens, and vitreous of a
bovine eye, as well as in human lens nuclei with varying
degrees of cataract. The research followed the tenets
of the Declaration of Helsinki. All thresholds were
measured in the experimental configurations used for
the investigation of the tissue effects described below.
We measured the minimal pulse energy required to
obtain 100% probability of occurrence of plasma formation
in a sequence of at least 30 pulses. The threshold
criterion was observation of the plasma radiation
in a darkened room by means of the slit lamp microscope.
Investigation of Tissue Effects
All experiments in this study were performed in vitro.
This allows precise focusing of the laser pulses, an easy
determination of their damage range, and the photography
of the cavitation bubble dynamics within the
corneal stroma. Whole enucleated eyes were used
whenever possible, but for some experiments tissue
specimens were prepared to achieve the required aiming
precision or suitable conditions for photography.
Membrane cutting on Descemet 's membrane was used
as a model to compare the precision achievable with ps
and ns pulses. Corneal specimens were obtained from
freshly (2 to 4 hours) enucleated bovine eyes collected
water filled
glass cuvette
/ A contact lens
mirror/
beamsplitter
Nd:YAG
protection filter
Slit lamp
without
front lens
FIGURE l. Experimental arrangement for the investigation of tissue effects of picosecond and
nanosecond Nd:YAG laser pulses and for time-resolved photography of the cavitation bubble
dynamics within the corneal stroma.

Tissue Effects of Picosecond and Nanosecond Photodisruption 3035
from a slaughterhouse. A central portion of the cornea
was excised using a 10 mm trephine. Each specimen
was mounted on a teflon holder and immersed in
a cuvette with physiological saline. The specimen
could be precisely positioned with an xyz-micrometer
stage. To produce incisions in Descemet's membrane,
the laser pulses were applied from the endothelial side
with a repetition rate of 10 Hz, when the specimen was
moved laterally with the translation stage (Fig. 2a). We
compared the effects of ps pulses with 50 n] pulse
energy to those of ns pulses with a pulse energy of 1
mj. Both energies are close to the threshold for plasma
formation at the respective pulse duration.
Intrastromal corneal laser effects were investigated to
analyze the feasibility of intrastromal refractive surgery
with ps pulses. Enucleated bovine eyes (n = 6)
were mounted into an eye holder fixed on a translation
stage. Initially, the optical axis of the eye was coaxial to
the laser beam axis, and the laser pulses were focused
through the cornea onto the endothelium (see Fig. 2b,
which shows an enlarged view of the central cornea).
The eye was then moved laterally while ps pulses with
80 ^J pulse energy and 10 Hz repetition rate were
applied. The laser effects produced this way were located
at an increasing distance from the corneal endothelium
because of the curvature of the cornea. This
allows an accurate determination of the range for endothelial
damage from serial sections through the
laser effects.
To compare the suitability of ps and ns pulses for
cataract fragmentation, laser pulses were focused into
a)
FIGURE 2. Irradiation geometries for the investigation of tissue
effects produced by picosecond and nanosecond
Nd:YAG laser pulses, (a) Incision of Descemet's membrane
in a corneal specimen, (b) Corneal instrastromal tissue evaporation
in an enucleated eye (enlarged view), (c) Cataract
emulsification. (d) Focusing into the vitreous body close to
the retina.
the clear lenses of whole enucleated bovine eyes and
into human lens nuclei obtained by extracapsular cataract
extraction. One hundred microjoule ps pulses
and 1.1 mj ns pulses were focused into the bovine
lenses (n = 11) at a distance of about 4 mm behind the
anterior lens capsule (Fig. 2c), generating a series of
parallel lines. The laser effects were documented by
slit lamp photography. The human lens nuclei were
stored in saline at 4°C for a maximum of 24 hours
until the experiment was performed. The lenses (n =
23) had varying degrees and types of cataract. After
removing any cortical remnants, the lens nuclei were
gently clamped in a holder and immersed into a cuvette
filled with saline, and ps and ns laser pulses were
focused into the posterior third of the nuclei. The
threshold for plasma formation was determined as described
above in the clearest and in the most turbid
parts of the lens nuclei.
To measure the range for retinal damage during vitreous
surgery with ps pulses, we used the irradiation
geometry shown in Figure 2d. Bovine eyes (n = 7) were
hemidissected and mounted on an xyz-translation
stage, whereby traction on the retina was avoided as
much as possible. Series of ps laser pulses with a pulse
energy of 200 /xj each were then focused into the vitreous
at various well-defined distances from the internal
limiting membrane, where the retinal vessels are located.
The distance between the retinal vessels and the
laser focus was controlled with the micrometer translation
stage. The precise control of this distance was
facilitated by the eyecup preparation used, allowing
optimal illumination and observation of the retina.
Histology. After laser exposure, all corneal specimens
were fixed in 4% glutaraldehyde, postfixed in
buffered 2% osmium tetroxide, and dehydrated in alcohol.
The specimens with intrastromal laser effects
were put into propyleneoxide and then embedded in
epon 812. Series of semithin sections (1.5 ixm) were
stained with toluidine blue for light microscopy. The
specimens with incisions in Descemet's membrane
were critical point dried in CO2 and sputter coated
with gold. The surface morphology of the lesions was
studied using a scanning electron microscope (JSM
35, JEOL, Tokyo, Japan). After scanning electron microscopy,
the specimens were put into 100% ethanol,
then in propyleneoxide, and later they were embedded
in epon 812 for the preparation of semithin sections.
In this way, the surface morphology and histologic
sections of the same corneal lesions could be investigated.
For analysis of the laser effects in the retina, the
whole hemidissected globe was fixed in a 4% glutaraldehyde
solution for at least 24 hours. Afterward, most
of the vitreous was removed, and retinal specimens
containing the lesions were prepared. They were further
processed for histology as described above for the
corneal specimens.

3036 Investigative Ophthalmology 8c Visual Science, June 1994, Vol. 35, No. 7
TABLE l. Threshold Energies for Plasma Formation With ps- and
ns-Pulses in Clear Ocular Media of Bovine Eyes
Tissue
Cornea (stroma and
Descemet's membrane)
Lens
Vitreous
Threshold for
ps-pulses (nj)
15
48
35
Photography of Cavitation Bubble Dynamics
Within the Corneal Stroma
The cavitation bubble dynamics within the corneal
stroma were studied by means of flash photography at
different time delays between the optical breakdown
and the exposure of the film. The setup is shown in
Figure 1. The light source (Impulsphysik [Hamburg,
Germany], Nanolite KL-L) delivered spark flashes
with a duration of 20 ns. The light was collected by a
lens with large aperture (Canon, [Tokyo, Japan], F =
1.2, f = 50 mm) and collimated by a second lens (Pentax,
[Tokyo, Japan], F = 1.8, f = 50 mm). The photographs
were taken with 7X magnification on Kodak
(Rochester, NY) T Max 100 film using a Leitz (Wetzlar,
Germany) Photar lens (F = 3.5, f = 40 mm) attached
to the body of a 35 mm camera. The laser pulse
and the flash were released while the camera shutter
was open and the room light darkened. To trigger the
flash lamp, a small part of the Nd:YAG laser light was
directed onto a fast photodiode connected to a delay
generator. The time between the laser pulse and the
flash could be adjusted in steps of 1 ;us starting from 2
/is to any longer delay required.
Cornea specimens from sheep eyes obtained from
a slaughterhouse (n = 15) were mounted on a teflon
holder and immersed into the cuvette filled with physiological
saline. To get smooth plane cuts, the corneal
excisions were performed with a preparation blade.
Picosecond laser pulses with 80 ft] and 300 /x] pulse
energy were focused through the contact lens into the
corneal stroma. The light was incident from the epithelial
side. Photographs were taken through a side of the
corneal specimen such that the pictures showed a
cross-sectional view of the cornea displaying the location
of the cavities with respect to the epithelium and
the endothelium. To minimize edematous changes of
the corneal stroma, the experiments were performed
within 10 minutes of excision of the specimen. For
comparison, the bubble dynamics in water were also
documented.
RESULTS
Threshold Energies for Plasma Formation
Table 1 presents the threshold values of the energy
required for plasma formation in cornea, lens, and
Threshold for
ns-pulses (nj)
180
560
445
Threshold Ratio
(ns/ps)
12.0
11.6
12.7
vitreous of bovine eyes, and the ratios of the threshold
values for ns and ps pulses. On average, 12 times less
energy is required to produce a plasma with ps pulses
than with ns pulses. Figure 3 compares the threshold
energies for ps and ns laser pulses focused into the
posterior third of cataractous human lens nuclei (n =
23). The threshold energies varied considerably both
between the individual nuclei and depending on
whether the laser focus was located in clear or opaque
parts of the lenses. Therefore, the lowest and highest
threshold values observed with both pulse durations
are indicated for each lens. They are connected by a
line to demonstrate which values belong to the same
lens. With ns pulses, an energy of up to 21 mj was
required to produce a plasma, whereas with ps pulses
an energy of 3 mj was always sufficient. On average,
0.0 0.5 1.0 1.5 2.0 2.5 3.0
ps - laser pulse energy [mJ]
FIGURE 3. Comparison of breakdown thresholds for ps and
ns pulses focused into the posterior third of cataractous human
lens nuclei (n = 23). For each lens, the lowest and highest
threshold values for both pulse durations observed in the
clearest parts of the nucleus and those with the most scattering
are plotted as dots. Each pair of dots is connected with a
line to show which values belong to the same lens. Note the
different scales for the ps and ns pulse energies.

Tissue Effects of Picosecond and Nanosecond Photodisruption
FIGURE 4. (a, b) Scanning electron micrographs of cuts through Descemet's membrane. The
cuts were produced with 50 jxj ps pulses in (a) and with 1 mj ns pulses in (b). In each case,
about 100 pulses were applied at a repetition rate of 10 Hz. (c, d) Semithin sections through
the lesions shown in (a, b). The length of the scales corresponds to 100 ftm.
the pulse energy could be reduced by a factor of 7 6 to 8. Figure 6 shows plasma formation and cavitation
when ps pulses were used instead of ns pulses. in corneal tissue in comparison to the maximal bubble
Tissue Effects on Descemet's Membrane
size reached in water. Both the effects in cornea and in
water were produced with 300 /uj pulse energy. The
Figure 4 shows scanning electron micrographs and se- bubbles in water expand rapidly in about 40 to 45 /us
mithin sections of cuts in Descemet's membrane producedand
reach the collapsed stage after 80 to 90 us. The
with ps and ns Nd:YAG laser pulses. Using ps pulses bubble expansion in the cornea is strongly decelerated
with an energy of 50 fx] per pulse, it was possible to because of the high viscosity of the corneal tissue. The
achieve a dissection of Descemet's membrane with a photograph in Figure 6, taken 3 /is after breakdown,
width of 30 to 40 /im and a range of endothelial dam- shows the maximally expanded stage of the bubble in
age of about 130 /xm (Figs. 4a, 4c). The width of the
dissection is only slightly larger than the plasma diame-
the cornea. It is much smaller than the bubble in water
(it has less than
ter. Nanosecond pulses with a pulse energy of 1 mj,
however, caused gross ruptures in Descemet's membrane
and a large collateral damage zone of about 400
/um (Figs. 4b, 4d). A dislocation of the stromal lamellae
below the cut is observed in both cases, but it is
stronger for the ns pulses. Figure 5 demonstrates that
it is possible to produce laser effects that are almost
selective to the corneal endothelium. To achieve this
high degree of spatial confinement, ps pulses with an
energy of 35 /ij (i.e., close to the breakdown threshold)
were used.
Cavitation and Tissue Effects in the Corneal
Stroma
Intrastromal corneal tissue effects produced by ps pulses
and their origin via cavitation are presented in Figures
] /3 of its diameter and about ] /3o part of
its volume). Although the bubble expansion ceases in
less than 3 n& after plasma formation, the bubbles
reach a size much larger than the volume of the laser
plasma (see Fig. 6). Even 2 minutes after breakdown,
the volume of the cavity is still more than 10 times
larger than the plasma volume. This demonstrates that
the effect of the laser pulses is mainly displacement but
only to a small extent evaporation of tissue. The intrastromal
cavities collapse slowly and disappear only
about 1 hour after their generation. The lobular form
of the bubbles immediately after their generation is
probably related to the arrangement of the collagen
lamellae within the stroma.
Figure 7 presents the histologic appearance of intrastromal
laser effects created with 80 ti] ps pulses.
The corneal specimens were fixed about 5 minutes
3037

3038 Investigative Ophthalmology 8c Visual Science, June 1994, Vol. 35, No. 7
FIGURE 5. (a) Scanning electron micrograph of laser effects
selectively produced in the corneal endothelium using a series
of ps pulses with a pulse energy of 35 juj each, (b) Seinithin
section through the same lesion. The scales correspond
to a length of 100 nm.
after laser exposure. At this time, the cavities had a
diameter of about 60 /mi, and the cavity volume exceeded
the volume of evaporated tissue, which is approximately
equal to the plasma volume by a factor of
more than 10. Correspondingly, a strong deformation
of the collagen lamellae around the cavity is observed.
In Figures 7a and 7b, some endothelial cells are missing
below the laser effects. Damage detectable by light
microscopy was observed up to a distance of 150 pm
between the laser focus and the corneal endothelium.
Figure 8 shows intrastromal effects created when
the laser pulses were applied with 10 Hz repetition
rate while the specimen was slowly moved in a lateral
direction. Using a pulse energy of 80 ^J, a homogeneous
cavity could be produced that has a width of 160
to 230 ^m and a length of 2 mm (Fig. 8a). The corneal
epithelium and Descemet's membrane are apparently
not damaged. An increase of the laser pulse energy to
300 fx] leads to a spongelike agglomeration of bubbles
rather than to one homogeneous cavity (Fig. 8b). Series
of individual cavities were, however, observed as
well with 80 ^J pulse energy.
Tissue Effects in the Lens
Figure 9 shows slit lamp photographs of ps and ns laser
effects in bovine lenses produced with 100 fij and 1.1 mj
pulse energy, respectively. The effects look similar to
the cavities produced in corneal tissue, and, like them,
they disappear after about 1 hour. The application of
1.1 mj ns pulses sometimes resulted in a splitting of
the lens fibers and a creation of one or several large
bubbles reaching up to the anterior lens capsule (Fig.
9b). These bubbles block the laser light path and the
view toward the posterior part of the lens. The ps laser
effects are much finer (the diameter of the cavities is
reduced approximately by a factor of 3), and a splitting
of the lens fibers was not observed.
The laser effects in human lens nuclei varied with
the optical and mechanical properties of the cataract.
Soft lenses developed vacuoles and bubbles where an
optical breakdown occurred, similar to the picture in
Figure 9. Dense nuclei developed more discrete pits,
cracks, and lamellar clefts independent of the pulse
length. The breakdown thresholds for both pulse durations
are given in Figure 3.
Tissue Effects at the Retina
Figure 10 is a macrophotograph of laser effects that
were produced with a series of 200 n] ps pulses to
determine the range for retinal damage during vitreous
surgery close to the vitreoretinal interface. In the
center of each line of laser exposures, the laser focus
was located in the vitreous body. The distance Ax from
the retinal vessels varied between 200 jum and 800 ftm.
At a certain distance from the center, the laser focus
was located within the retina and further laterally it
2 min
SE
f
»
0.5 mm
FIGURE 6. Comparison between the cavitation bubble dynamics
in corneal stroma {left) and in water (right). The bubbles
in the cornea were photographed 3 fis, 10 seconds, and
2 minutes after the laser pulse was released, and the picture
of the bubble in water was taken 40 ^s after the laser pulse.
Both pictures are printed with the same magnification. All
bubbles were produced with 300 n] ps pulses.

Tissue Effects of Picosecond and Nanosecond Photodisruption 3039
b
FIGURE 7. Semithin sections through picosecond laser effects
within the corneal stroma. The pulse energy was 80 jtj.
The distance between laser focus and corneal endothelium
was 130 nn\ in (a) and 60 fim in (b). Endothelial damage is
marked by arrowheads. The scales correspond to a length of
100/ini.
was located at the level of the retinal pigment epithelium.
Focusing above a retinal vessel led to a hemorrhage
when the distance between the vessel and the
laser focus was less than 300 fj.m. When the pulses were
focused onto the retinal pigment epithelium or slightly
behind it, no macroscopic vessel damage could be observed,
even when the laser focus was located directly
below a vessel (arrow in Fig. 10). Figure 11 shows histo-
logic sections through retinal lesions arising when the
distance Ax between the laser focus in the vitreous and
the vitreoretinal interface was 200 ^m (Fig. 11a) and
400 fim (Fig. lib), respectively. In both cases, damage
to the neuroretina and the outer nuclear layer, probably
caused by the cavitation bubble dynamics, is visible.
For Ax > 500 /mi, no retinal damage was observed.
This means that the damage range is about half a millimeter
at a laser pulse energy of 200 /*J.
DISCUSSION
Thresholds for Plasma Formation
Table 1 shows that the threshold energy for plasma
formation in clear ocular media is, on average, 12
times lower for ps pulses than for ns pulses. This
agrees quite well with the factor of 13 reported by us
for distilled water. 6 Docchio et al 9 presented similar
values: a factor of 10 for calf vitreous and of 14 for
distilled water. The variation in the absolute values of
the breakdown thresholds for different tissues observed
at constant pulse duration can partly be attributed
to different tissue properties, and to some extent
to the different experimental configurations used for
the investigation of the various tissues (see Fig. 2).
The breakdown thresholds in cataractous lens nuclei
(Fig. 3) are always independent of the pulse dura-
FIGURE 8. Intrastromal cavities produced by a series of ps
pulses. The laser light was incident from the right. The pulse
energy was 80 /xj in (a) and 300 n] in (b). The pulses were
applied with 10 Hz repetition rate when the corneal specimen
was moved in a vertical direction. The scale represents
0.5 mm. The photographs were taken immediately (

3040 Investigative Ophthalmology & Visual Science, June 1994, Vol. 35, No. 7
FIGURE 9. Slit lamp photographs of Nd:YAG laser effects in
bovine lenses produced with 100 /uj ps pulses (a) and 1.1 mj
ns pulses (b). The laser pulses were applied with a 10 Hz
repetition rate while the lens specimen was moved in a horizontal
direction to create a pattern of parallel lines. The
laser focus was located about 4 mm behind the anterior lens
capsule. Besides the intended emulsification, the ns pulses
have caused the formation of a big bubble (arrowheads)
reaching up to the anterior lens capsule.
tion, much higher than in clear ocular media, because
the scattering of the laser light within the lens nucleus
leads to an enlargement and distortion of the focal
spot. The pulse energy needed for plasma formation
could, on average, be reduced by a factor of 7 when ps
pulses were used. It is so far not well understood why
this value is smaller than the energy reduction observed
in clear ocular media and water.
Precision of ps and ns Tissue Effects
The lower thresholds for plasma formation with ps
pulses are the basis for considerable improvement in
the precision of intraocular microsurgery. This is most
clearly demonstrated in Figure 4 showing cuts in Descemet's
membrane and in Figure 5 showing the re-
moval of a line of endothelial cells from Descemet's
membrane. It should be noted, however, that a reduction
of the pulse energy from 1 mj to 50 juj (i.e., by a
factor of 20) led to a decrease of the damage zone
from 400 jam to 130 pm (i.e., only by a factor of 3).
This result is in accordance with the earlier finding
that the damage range scales with the cube root of the
laser pulse energy. 7 ' 8 On the other hand, Figure 4 demonstrates
that not only the range but also the severity
of the damage is reduced with the lower pulse energy.
Similar conclusions can be drawn from the laser effects
within the lens (Fig. 9), where a splitting of the
lens fibers was only observed after ns pulses with an
energy of 1.1 mj. The range for endothelial damage
during corneal intrastromal surgery resembles that
observed after dissection of Descemet's membrane: It
is 150 ^m for ps pulses with 80 MJ pulse energy. The
range for retinal damage was larger: It was 500 fim
when 200 ^tj ps pulses were focused into the vitreous.
This may be partly due to the higher pulse energy used
and to some extent to the different tissue properties of
neuroretina and corneal endothelium.
Disruptive Action of Cavitation
Even with very small pulse energies, it is not possible to
eliminate disruptive effects entirely because intraocular
plasma formation is inevitably followed by a rapid
expansion of the plasma and, thus, by shock wave
emission and cavitation. The cavitation bubble dynamics
can cause considerable tissue displacement because
the bubble reaches a size much larger than the
laser plasma (see Fig. 6), and its initial expansion velocity
is several hundred meters per second. 6 ' 7 The effects
of the bubble expansion are most pronounced when
the plasma is formed within a tissue, for example, inside
the cornea or the lens. In the cornea, a cavity
more than 10 times as large as the plasma volume is
formed (see Figs. 6, 7) and disappears more than 1
hour after laser exposure. Besides resulting in a dislocation
of the stromal lamellae, the bubble expansion
can cause stress on the corneal endothelium and contribute
to the endothelial damage observed below the
cavities.
When the laser effects are produced in a fluid at
or close to a tissue surface, the bubble reaches a diameter
larger than that within the cornea or lens (see Fig.
6), but it can affect the tissue only from the liquid-tissue
interface. Therefore, the tissue displacement and
disruption are less pronounced than they are after focusing
laser pulses into the tissue. Examples for the
effects arising when the laser focus is located at the
tissue surface are the changes in Descemet's membrane
and the corneal endothelium shown in Figure 4
and the damage to the neuroretina visible in Figure
11. The disruptive effects are now only partly caused

Tissue Effects of Picosecond and Nanosecond Photodisruption 3041
800jinr
200|j,nr
Ax
400|anr
600[iri
RPE
location of
laser focus
FIGURE 10. Macrophotograph of laser effects within the retina caused by ps pulses with a
pulse energy of 200 n]. Ax is the distance between the laser focus in the vitreous and the
retinal vessels as adjusted in the cenLer of each line of laser exposures. At the sides of each
line, the laser focus was located within the retina or at the pigment epithelium. Focusing
directly onto a vessel always caused a hemorrhage. The arrow indicates a vessel showing no
macroscopic damage when the laser pulses had been focused on the pigment epithelium, that
is, at a level about-300 fim below the vessel.
by the bubble expansion and partly by the jet formation
during bubble collapse, which is described elsewhere.
714 Although the extent of tissue disruption by
cavitation varies from case to case and decreases when
the pulse energy is reduced, it is apparently the largest
obstacle to a spatial confinement of the effects of intraocular
photodisruption.
Potential Applications of Picosecond Pulses
Picosecond laser pulses can, in principle, be applied
for all indications for which nanosecond pulses are
presently used, whereby they offer the advantage of a
reduced damage range. When very fine tissue effects
are intended or when the application site is very close
to sensitive ocular structures, such as the corneal endothelium
or the retina, picosecond pulses are likely to
be the only means to achieve the surgical aim. Possible
new applications of this kind are corneal intrastromal
refractive surgery, cataract fragmentation, and vitreous
surgery close to the retina. Other uses could be
trabeculopuncture for treatment of glaucoma and, as
Figure 5 shows, the selective removal of a cell layer on
a tissue surface, for example, polishing a lens capsule
after cataract surgery.
Intrastromal refractive surgery. Several authors have
put forward the idea that refractive changes of the
cornea may be achieved by removal of an intrastromal
tissue layer through plasma-mediated evaporation.
1115 " 17 They have expressed the hope that the corneal
curvature can be changed without damaging the
corneal epithelium, Bowman's membrane, or the endothelium,
and that thereby the haze and regression
arising during the healing process of the cornea can be
diminished. This aim cannot be accomplished using
conventional photodisruptors delivering nanosecond
laser pulses because their tissue effects are too
coarse. 715 If, as an alternative, the ps laser is used for

3042 Investigative Ophthalmology &: Visual Science, June 1994, Vol. 35, No. 7
FIGURE 11. Histologic sections through retinal lesions arising
from 200 ^tj ps pulses focused into the vitreous at a distance
of 200 pm (a) and 400 /im (b) from the retina. The sections
have been obtained from the specimen shown in Figure 11.
refractive surgery, a key factor is to avoid endothelial
damage. Our experiments show that endothelial damage
occurs when 80 n] ps pulses are focused up to 130
fxm from the endothelium (Fig. 7). Damage was
avoided when the distance between endothelium and
laser focus was larger than 150 /xm. Fortunately, intrastromal
lamellar keratectomy will give the largest
change in corneal curvature if performed within the
anterior part of the cornea. 16 There are nevertheless
difficulties with this technique. Figures 6 to 8 demonstrate
that the volume of the evaporated tissue (which
approximately equals the plasma volume) is small compared
to the tissue displacement caused by the expansion
of the laser plasma. The laser-induced cavities disturb
the initial optical properties of the cornea, making
accurate focusing difficult. The bubbles from
earlier pulses reflect and refract the light of the subsequent
laser pulses. This can lead to the formation of
multiple plasmas and, thus, to multiple cavities, especially
when the laser pulse energy is relatively high.
The cavities are often distributed in an irregular manner
(Fig. 8), and it is not always possible to evaporate a
homogenous tissue layer of definite thickness. This re-
duces the likelihood of obtaining a predictable change
in corneal curvature.
Cataract fragmentation. The picosecond Nd:YAG
laser has been suggested as a method to fragment and
then aspirate a cataract through a small capsular
opening while retaining the lens capsule. 16 This allows
filling of the capsular bag with a clear material to restore
lenticular function. 18 " 20 Ultrasonic phacoemulsification
is difficult to use with hard nuclear cataracts
and requires a 2.5 to 3 mm incision. 21 Both the ns laser
and the mode-locked ps pulse trains need up to 10 mj
to fragment dense cataracts, 22 " 25 and even at these
high energies the probability of breakdown is still only
40%. 24 We observed that ns pulses up to 21 mj pulse
energy were required to produce a plasma in the posterior
third of cataractous human lenses (Fig. 3). At such
high pulse energies, there is a substantial risk of capsule
rupture from cavitation bubble expansion, even
when the laser pulses are focused well away from the
capsule. The use of single ps pulses of much lower
energy offers a solution to this problem. By ps pulses
of only 3 mj, plasmas were produced in any part of all
nuclei tested, and less than 1 mj was needed for 48%
of the 23 lens nuclei investigated. Because the fragmentation
of lens material requires a large number of
laser exposures, a high repetition rate of about 1 kHz,
together with a scanning system, would be necessary to
allow application of the laser exposures in a reasonable
time.
Vitreoretinal surgery close to the retina. Various clinicians
have reported on vitreoretinal surgery with ns
Nd:YAG laser pulses 426 " 32 and with mode-locked ps
pulse trains. 31 They have observed that when applying
pulse energies of 3 to 5 mj, there has to be a distance
of about 3 mm between the application site and the
retina to avoid retinal and choroidal hemorrhage.
4 ' 2G - 28 ' 31 This result agrees well with experimental
studies on rabbit eyes. 33 " 33 By employing ps pulses,
it will be possible to treat disorders much closer to the
retina because the retinal damage range for ps pulses
with 200 /uj pulse energy was found to be only 0.5 mm
(Figs. 10, 11). Near the optical axis of the eye, the
damage range may be even further reduced by using
pulse energies closer to the breakdown threshold of
about 35 /xj. A minimal damage range of approximately
200 juj seems to be achievable, considering that
the range is proportional to the cube root of the laser
pulse energy. 7 ' 8 This may render epiretinal membranotomy
possible, as well as the dissection of vitreous
strands in the vicinity of the retina. In the periphery,
the focal spot is degraded by aberrations caused by the
oblique passage of the laser light through the optical
system of the eye. This leads to an increase of the
threshold energy for plasma formation and establishes
a need to use higher pulse energies. This problem is,

Tissue Effects of Picosecond and Nanosecond Photodisruption 3043
however, the same for ns pulses so that the use of ps
pulses remains as advantageous in the periphery as it
does close to the optical axis.
Although an automatic scanning and aiming system
operating at high repetition rates would be useful
for phacofragmentation or corneal refractive surgery,
manual aiming seems to be most appropriate for vitreoretinal
laser applications. With manual aiming
controlled by the direct feedback from observing the
action of the laser pulses, the highest possible surgical
precision can be realized, together with a minimization
of the light energy deposited. To keep the total
amount of light energy applied as low as possible, only
moderate repetition rates of 10 to 100 Hz should be
used. Light absorption in the retinal pigment epithelium
might otherwise lead to heating and coagulation
of the retina, in addition to possible mechanical damage
associated with misaimed laser pulses.
CONCLUSIONS
The use of ps pulses with a moderate repetition rate
(10 Hz to 1 kHz) and energies in the microjoule range
is a new concept compared to the well-established
Nd:YAG laser surgery with single ns pulses. It increases
the surgical precision and reduces the disruptive
side effects in "classical" Nd:YAG laser applications
like posterior capsulotomy and iridotomy and in
the whole range of other applications for which treatment
with ns pulses is already well established. Besides
that, it also renders new applications possible. Our
preliminary investigations suggest that cataract emulsification,
vitreoretinal surgery close to the retina, and
intrastromal corneal refractive surgery deserve further
in vivo experiments and clinical studies. An instrument
with a repetition rate variable between 10 Hz
and about 1 kHz, offering the possibility of manual
aiming and automatic scanning, would be most versatile.
Pulse energies below 1 mj will be sufficient in
most cases, but for cataract fragmentation the range
of pulse energies available should reach up to about 2
to 3 mj to allow fragmentation of dense cataracts.
Key Words
Nd:YAG laser, picosecond pulses, refractive surgery, cataract
fragmentation, vitreoretinal surgery
Acknowledgments
The authors thank M. Volkholz, C. Grosse, and U. Weinhardt
for their support during the histologic investigations,
and L. Merz, H. Krohn, R. Kube, and R. Carbe for their help
in preparing the manuscript.
References
1. 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. Int Ophthahnol.
1982;5:15-32.
2. Aron-Rosa D, Aron J, Griesemann J, Thyzel R. Use of
the neodym YAG laser to open the posterior capsule
after lens implant surgery: A preliminary report./ Am
Intraocul Implant Soc. 1980;6:352-354.
3. Steinert RF, Puliafito CA. The Nd:YAG Laser in Ophthalmology.
Philadelphia: WB Saunders; 1985.
4. Fankhauser F, Kwasniewska S. Neodymium:yttriumaluminium-garnet
laser. In: L'Esperance FA, ed. Ophthalmic
Lasers.' 3rd ed. St. Louis: CV Mosby;
1989:781-886.
5. Barnes PA, Rieckhoff KE. Laser induced underwater
sparks. Appl Phys Lett. 1968;13:282-284.
6. Vogel A, Busch S, Jungnickel K, Birngruber R. Mechanisms
of intraocular photodisruption with picosecond
and nanosecond laser pulses. Lasers Surg Med.
1994;14:(in press).
7. Vogel A, Schweiger P, Frieser A, Asiyo MN, Birngruber
R. Intraocular Nd:YAG laser surgery: Lighttissue
interaction, damage range, and reduction of collateral
effects. IEEE J Quant Electr. 1990;26:2240-
2260.
8. Zysset B, Fujimoto JG, Puliafito CA, Birngruber R,
Deutsch TF. Picosecond optical breakdown: Tissue effects
and reduction of collateral damage. Lasers Surg
Med. 1989;9:193-204.
9. Docchio F, Sacchi CA, Marshall J. Experimental investigation
of optical breakdown thresholds in ocular media
under single pulse irradiation with different pulse
durations. Lasers Ophthahnol. 1986;l:83-93.
10. Zysset B, Fujimoto G, Deutsch TF. Time resolved measurements
of picosecond optical breakdown. Appl
PhysB. 1989;48:139-147.
11. Niemz MH, Hoppeler TP, Juhasz T, Bille JF. Intrastromal
ablations for refractive corneal surgery using
picosecond infrared laser pulses. Lasers Light Ophthalmol.
1993;5:149-155.
12. Vogel A, Hentschel W, Holzfuss J, Lauterborn W. Cavitation
bubble dynamics and acoustic transient generation
in ocular surgery with pulsed neodymium: YAG
lasers. Ophthalmology. 1986;93:1259-1269.
13. Mellerio J, Capon M, Docchio F. Nd:YAG lasers: A
potential hazard from cavitation bubble behaviour in
anterior chamber procedures? Lasers Ophthahnol.
1987;l:185-190.
14. Vogel A, Lauterborn W, Timm R. Optical and acoustic
investigations of the dynamics of laser-produced
cavitation bubbles near a solid boundary. J Fluid Mech.
1989;206:299-338.
15. Hoh H, Becker KW. Intrastromal keratohexis with
the Nd:YAG laser:. A possible way of refractive
surgery? Klin Mbl Augenheilk. 1990;197:480-487. In
German.
16. Bille JF, Klancnik EG, Niemz MH. Principles of operation
and first clinical results using the picosecond IR
laser. In: Parel JM, ed. Ophthalmic Technologies II.
SPIEProc. 1992;1644:88-95.

3044 Investigative Ophthalmology & Visual Science, June 1994, Vol. 35, No. 7
17. Tabaoda J. Micron-sized laser supported plasma effects
applied to microsurgery. In: Parel JM, ed. Ophthalmic
Technologies II. SPIE Proc. 1992;1644:72-78.
18. Kessler J. Experiments in refilling the lens. Arch Ophthalmol.
1964;71:412-417.
19. Parel JM, Gelender H, Trefers WF, Norton EWD.
Phaco-Ersatz: Cataract surgery designed to preserve
accommodation. Graefes Arch Clin Exp Ophthalmol.
1986;224:165-173.
20. Haelliger E, Parel JM, Fantes F, et al. Accommodation
of an endocapsular silicone lens (Phaco-Ersatz) in
the nonhuman primate. Ophthalmology. 1987;94:471-
477.
21. Emery JM, Mclntyre DJ. Extracapsular Cataract Surgery.
St. Louis: CV Mosby; 1983.
22. Zelman J. Photophaco fragmentation. J Cataract Refract
Surg. 1987;13:287-289.
23. Chambless WS. NeodymiunrYAG laser phacofracture:
An aid to phacoemulsification. J Cataract Refract
Surg. 1988;14:180-181.
24. Ryan EH, Logani S. Nd:YAG laser photodisruption
for the lens nucleus before phacoemulsification. Am]
Ophthalmol. 1987; 104:382-386.
25. Flock BW, Chew PTK, Lim ASM, Tock EPC. Qswitched
Nd:YAG laser disruption of rabbit lens nucleus.
Laser Light Ophthalmol. 1990;3:227-232.
26. Fankhauser F, Loertscher HP, van der Zypen E. Clinical
studies on high and low power laser radiation upon
some structures of the anterior and posterior segment
of the eye. hit Ophthalmol. 1982;5:15-32.
27. Fankhauser F, Kwasniewska S, van der Zypen E. Vitreolysis
with the Q-switched laser. Arch Ophthalmol.
1985;1O3:1166-1171.
28. Brown GC, Benson WE. Treatment of diabetic traction
retinal detachment with the pulsed Neodymium-
YAG laser. Am] Ophthalmol. 1985;99:258-262.
29. Little HL, Jack PL. Q-switched Neodymium-YAG
laser surgery of the vitreous. Graefes Arch Clin Exp Ophthalmol.
1986;224:240-246.
30. Jagger JD, Hamilton P, Polkinghorne P. Q-switched
neodymium YAG laser vitreolysis in the therapy of posterior
segment disease. Graefes Arch Clin Exp Ophthalmol.
1990;228:222-225.
31. Tassignon MJ, Kreissig I, Strempels N, Brihave M. Indications
for Q-switched and mode-locked Nd:YAG
lasers in vitreoretinal pathology. Eur J Ophthalmol.
1991:1:123-130.
32. Gabel VP. Lasers in vitreoretinal therapy. Ann Ophthalmic
Laser Surg. 1992:1:75-89.
33. Hiemer H, Hutter H, Gabel VP, Birngruber R. Retinal
damage by the Neodymium-YAG laser application
in the vitreous. Fortschr Ophthalmol. 1985;82:447-
449. In German.
34. Bonner RF, Meyers SM, Gaasterland DE. Threshold
for retinal damage associated with the use of highpower
neodymium-YAG lasers in the vitreous. Am]
Ophthalmol. 1983;96:153-159.
35. Puliafito CA, Wasson PJ, Steinert RF, Gragoudas E.
Neodymium-YAG surgery on experimental vitreous
membranes. Arch Ophthalmol. 1984;102:843-847: