Picosecond acoustics in vegetal cells: Non-invasive in vitro measurements at a sub-cell scale

Ultrasonics 50 (2010) 202–207

Contents lists available at ScienceDirect

Ultrasonics journal homepage: www.elsevier.com/locate/ultras

Picosecond acoustics in vegetal cells: Non-invasive in vitro measurements at a sub-cell scale B. Audoin a,*, C. Rossignol a, N. Chigarev a, M. Ducousso a, G. Forget b, F. Guillemot b, M.C. Durrieu b a b

Université de Bordeaux, CNRS, UMR 5469, Talence F-33405, France Université de Bordeaux, INSERM, UMR 577, Bordeaux F-33000, France

a r t i c l e

i n f o

Article history: Received 22 July 2009 Received in revised form 11 September 2009 Accepted 14 September 2009 Available online 27 September 2009 PACS: 43.35 43.80 43.20 Keywords: Acoustic microscopy Biophononics Cell imaging Medical applications Picosecond acoustics

a b s t r a c t A 100 fs laser pulse passes through a single transparent cell and is absorbed at the surface of a metallic substrate. Picosecond acoustic waves are generated and propagate through the cell in contact with the metal. Interaction of the high frequency acoustic pulse with a probe laser light gives rise to Brillouin oscillations. The measurements are thus made with lasers for both the opto-acoustic generation and the acousto-optic detection, and acoustic frequencies as high as 11 GHz can be detected, as reported in this paper. The technique offers perspectives for single cell imaging. The in-plane resolution is limited by the pump and probe spot sizes, i.e. 1 lm, and the in-depth resolution is provided by the acoustic frequencies, typically in the GHz range. The effect of the technique on cell safety is discussed. Experiments achieved in vegetal cells illustrate the reproducibility and sensitivity of the measurements. The acoustic responses of cell organelles are significantly different. The results support the potentialities of the hypersonic non-invasive technique in the fields of bio-engineering and medicine. Ó 2009 Elsevier B.V. All rights reserved.

1. Introduction The laser–ultrasonic techniques, in which ultrasonic waves are both generated and detected with laser light, are receiving growing attention because of their wide range of applicability in nondestructive control and evaluation. These techniques are particularly well suited for material characterisation when measurements must be made without contacting or applying coupling gel to the structure. Moreover, since the duration of the laser pulse can be very short, high resolution in space (nm in-depth) and time (ps) can be achieved. Femtosecond laser pulses have been used, since the end of the 80s, to perform generation and detection of longitudinal acoustic waves in sub-micrometric films, multilayers structures and other nanostructures [1]. Ultrafast ultrasonic technique consists in measuring the transient reflectivity changes induced in a structure by the propagation of an optically generated acoustic wave. The absorption of a light pulse, namely the pump pulse, sets up a local

* Corresponding author. Address: Laboratoire de Mécanique Physique, Université Bordeaux 1, 351 Cours de la Libération, F-33405 Talence, France. Tel.: +33 5 40 00 69 69; fax: +33 5 40 00 69 64. E-mail address: [email protected] (B. Audoin). 0041-624X/$ - see front matter Ó 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.ultras.2009.09.019

thermal stress. This stress generates an elastic strain pulse propagating through the sample. The acoustic perturbation induces a change in the optical reflectivity of the structure. The reflectivity change is probed by a transient reflectometric or interferometric detection [2]. The main application fields of ultrafast opto-acoustics today are solid state physics and micro electronics. The latter is requiring an accurate mean of measuring the thicknesses and of controlling the bonding at a nanometer scale, that is provided by the picosecond acoustic technique. Despite recent development of femtosecond photothermal techniques in the field of biology, no application of picosecond acoustics has concerned this field up to now. However picosecond acoustics offers promising potentialities for biological imaging in the immediate future. Recently, measurement of high frequency acoustic waves generated with femtosecond lasers were performed by other teams in water [3] and liquid mercury [4]. We have recently reported on preliminary measurements (5–10 GHz) performed on vegetal cells [5]. The Section 2 of this paper presents the picosecond photoacoustic generation in a cell. Then a one dimension model, Section 3, accounting for photothermal generation, thermo-acoustic coupling and acousto-optic detection allows to predict the measured

203

B. Audoin et al. / Ultrasonics 50 (2010) 202–207

2. Opto-acoustic generation in a single cell Experiments mentioned in this paper are achieved with allium cepa (common onion) cells. Slices of alive cells are deposited directly on the surface of a polished titanium alloy (Ti6Al4 V) substrate. The thickness and the lateral size of a cell are measured by optical microscopy as approximately 5 lm and 50–100 lm, respectively. All the experiments are performed at the room temperature of 21 °C without any treatment or fixation of the alive cells. Well known reflectometric pump–probe experimental set-up is used [1] to generate and detect acoustic waves in the cells. The radiation of a femtosecond laser (wavelength 790 nm, mean power 1 W, pulse duration 100 fs, energy of the pulse 10 nJ) is divided to provide the pump and probe beams. The pump beam is modulated at a frequency of 330 kHz to provide a reference signal for lock-in amplification. Pump pulses are frequency-doubled (395 nm) through a BBO crystal. In our experiments, the pump beam mean power is reduced to 50 lW, hence to an energy of 0.5 pJ per pulse. An optical delay line (0–12 ns) is introduced in the probe beam path. Intensity of the reflected probe beam is detected by a photodiode. Such a set-up allows the detection of reflectivity relative variations at a level as low as 107. Note that pump and probe wavelengths can be exchanged, thus providing either blue or red light for each of them. Both beams are focused at the cell–substrate interface by a same objective with a 50 magnification. The width at mid-height of the space cross-correlation of the pump and probe beams is approximately 2 lm. Absorption of the pump laser beam in the substrate provides a local change of the temperature, which induces thermal and acoustic strains. The so generated sound pulse propagates with longitudinal sound velocity in the cell and in the substrate. The space profiles of the acoustic pulses are mainly determined by the optical penetration depth in the absorbing substrate, typically 10 nm for metals. The acoustic contribution to the reflectivity changes measured at a same location within a vegetal cell with either blue (k = 395 nm) or red (k = 790 nm) probe light are shown in Fig. 1. Oscillations, revealing acoustic contributions at 11 GHz and 5.5 GHz, respectively, are detected with a very good signal to noise ratio. The nature of these oscillations will be discussed in the following section.

= 395 nm

-6

relative intensity change (10 )

signals. It is used in Section 4 to comment on the maximum temperature rise in the cell and the maximum stress applied at the cell–substrate interface. Repetitive measurements, achieved on two species of a same cell variety, illustrate the sensitivity and the reproducibility of the measurements. Finally, the potentialities for cell imaging with a micron lateral resolution are shown in Section 6 of this paper.

2 = 790 nm

0

400

600

800

1000

1200

time (ps) Fig. 1. Acoustic signals measured at a same position within a cell for probe light at k = 395 nm (top), and k = 790 nm (bottom).



Q 1 ðz; tÞ ¼ 0 Q 2 ðz; tÞ ¼ b2 I2 dðtÞeb2 z

ð1Þ

where b2 stands for the optical absorption coefficient, I2 = I1(1  R12) with R12 the reflection coefficient at the cell–substrate interface and I1 the laser pulse fluence in the cell. Owing to the high transmission at the cell surface, I1 can be considered equal to the incident laser fluence per pulse. The temperature fields Ti(z, t) comply with the heat diffusion equations

qi C pi

@T i ðz; tÞ @ 2 T i ðz; tÞ ¼ Q i ðz; tÞ  ji @t @z2

ð2Þ

where qi, Cpi and ji denote the mass densities, specific heat capacities and thermal conductivities, respectively. Since no heat source exists in the cell, the temperature rise in this medium results of the temperature and heat flux continuities at the cell–substrate interface. For a correct representation of the interface, thermal resistivity is introduced in the boundary equation as

T 2 ð0; tÞ  T 1 ð0; tÞ ¼ RUi

ð3Þ

with Ui the heat flux at the interface. The sudden heating of the media generates transient acoustic displacements ui(z, t). They are solutions of the wave propagation equations

Ci

@ 2 ui ðz; tÞ @ 2 ui ðz; tÞ @T i ðz; tÞ  qi ¼ ki 2 @z @z @t 2

ð4Þ

where Ci and ki are the stiffness coefficients and the thermal rigidity coefficients, respectively. To account for sound attenuation in the media complex stiffness coefficients are considered. The displacements are such that continuity of the stress

3. In silico

ri ðz; tÞ ¼ C i In this section, the opto-acoustic wave generation and detection in the cell are simulated. The model couples the temperature diffusion equation and the wave motion equation. Finally, the reflectivity changes are calculated accounting for both the temperature rise and the acoustic strain. The cell and substrate are modeled as two infinite half spaces in contact at the plane z = 0, with the z axis oriented toward the depth of the substrate. In the following, index i = 1, 2 refers to the cell (z < 0) and to the substrate (z > 0), respectively. Assuming no light absorption in the cell, the heat sources brought by light absorption in the media are

200

@ui ðz; tÞ  ki T i ðz; tÞ @z

ð5Þ

is satisfied at the interface (z = 0). At this stage, one should remind that the physical data of experimental access are the relative changes of the probe light intensity. In the following, the acousto-optic interaction in the substrate is neglected. Let f(z) denote the intensity change caused by a perturbation of the optical index at position z. The principle of the optical detection of the strain pulse is shown in Fig. 2. Owing to the acousto-optic interaction in the cell, a part of the electromagnetic wave is propagated backward to the detector (beam (a) in Fig. 2). The remaining electromagnetic energy is reflected at the cell–substrate

204

B. Audoin et al. / Ultrasonics 50 (2010) 202–207

fB ¼

Fig. 2. Optical detection of the strain pulse in a transparent cell.

interface. It interacts again with the acoustic front, then propagates to the receptor (beam (b) in Fig. 2). Interferences of the two electromagnetic waves are recorded on the photodiode. f(z) is the socalled sensitivity function, the calculation of which is not detailed in this text [6]. It accounts for both the optical index changes in the cell, and the motion of the cell–substrate interface. Noting I0 the initial value of the probe light intensity, the relative changes are related to the index perturbation, caused by both heat and strain by:

DIðtÞ ¼ I0

Z

0



 @n1 @n g1 ðz; tÞ þ 1 T 1 ðz; tÞ f ðzÞdz @ g1 @T 1

1

ð6Þ

where g1 = ou1/oz is the elastic strain in the cell, and on1/og1 and on1/oT1 are constant acousto-optic and thermo-optic coefficients, respectively. The acousto-optic and thermo-optic coupling within the metallic substrate are neglected in Eq. (6). Writing Eqs. (2)–(5) in a Fourier domain with respect to time allows linearization of the coupled heat diffusion and wave motion equations, and of the boundary equations as well. Explicit forms for the temperature and displacement fields in either medium are then derived, hence the relative changes of the intensity DI(t)/I0. The total relative changes in the intensity, recorded in the vacuole of an allium cepa cell, is shown in Fig. 3. The very high signal to noise ratio of the measured signal can be underlined. The signal is mainly composed of a slow decrease associated with heat diffusion and a quasi-monochromatic oscillation. The so-called Brillouin scattering oscillations are the result of the interferences of beams (a) and (b) already mentioned, and shown in Fig. 2. These Brillouin oscillations can also be seen as oscillations of light intensity in a dynamic Fabry–Perot cavity bounded by the acoustic front and by the cell–substrate interface. Their frequency at normal incidence of probe beam is given by [3]:

14

-6

relative intensity change (10 )

12 10 8

2n1 v 1 k

ð7Þ

where k is the optical wavelength, and v1 is the sound velocity. Note that the ratio of the oscillation frequencies (11 GHz and 5.5 GHz) measured at a same position for two probe wavelengths (395 nm and 790 nm), Fig. 1, confirms the Brillouin nature of the detected intensity changes. If the optical index n1 of the medium is known, it is straightforward to obtain the value of the sound velocity from relation (1). The coefficient of sound attenuation a1 at the frequency fB can also be evaluated from the attenuation time of the oscillations s as a1 = (v1s)1, and the oscillations amplitude and phase as well, see Sections 5 and 6. The signal calculated for the vacuole of the allium cepa cell is shown in Fig. 3. For these calculations, the optical index of the cell was fixed to a typical value for vegetal cells [7], and the mass density was approximated to that of water. Starting from the modelling in [8], the acousto-optic on1/og1 and thermo-optic on1/oT1 coefficients have been estimated to 0.4 and 7  103 K1, respectively. The thermal conductivity j1 and heat capacity Cp1 of the cell are chosen to match the sudden rise and slow decrease of the experimental signal. Then, suppressing the thermal background in the waveforms, a fitting process allows identification of the complex stiffness coefficient of the cell. The data considered for the calculations are shown in Table 1. The model can now be used to estimate physical data that are not of direct experimental access such as the temperature rise near the cell–substrate interface and the stress induced at the interface. 4. Cell safety At this point the question arises whether the experiments themselves may provoke any damage to the cell. In this section the effects of the laser radiation, the temperature rise and the stress are analyzed successively. 4.1. Laser irradiation Several biomedical applications in which femtosecond lasers are focused in a cell have already been developed. Femtosecond laser pulses have been used for two-photon optical imaging of tissues using photophores [9–11] without observing morphological changes of the cell. At a cell scale, and in this context of two-photon optical microscopy, the influence of femtosecond near infrared microirradiation has been specially studied [12]. Chinese hamster ovary cells have been irradiated with light pulses similar to those considered in this paper, in terms of beam focusing and pulse duration, wavelength and repetition rate. It was shown that vitality and cellular reproduction of the cells remain unaffected by the

Table 1 Physical constants considered in the simulations for cell, water and titanium. Data marked with an asterisk are identified with a fit of the experimental signal in the vacuole of an allium cepa cell. No value can be found in the literature for the thermal resistance. The expression, R = e/j where e and j are the thickness and thermal conductivity of the interface, yields values for R in the range R = 109 K W1 for an interface thickness of 1 nm.

6 4 2 0 0

200

400

600

800

1000

1200

time (ps) Fig. 3. Experimental and calculated (diamonds) signals in the vacuole of a vegetal cell for a probe wavelength of 790 nm.

Heat capacity Cpi [J kg1 K1] Thermal conductivity ji [W m1 K1] Density [m3] Optical absorption bi [m1] Optical index ni Stiffness coefficient Ci [GPa]

Vacuole

Water

Titanium

9460* 0.7*

4190 0.6

500 21.9

1000 – 1.4 + j4  103 3.1 + j0.3*

1000 23 1.33 + j4  103 2.2 + j0.1

4507 5.2  107 2.5 + j4.3 155

205

B. Audoin et al. / Ultrasonics 50 (2010) 202–207

femtosecond microbeam for mean power up to 1 mW. This threshold is by more than one order of magnitude larger than the mean laser power used in the experiments reported in this paper. Also, viability has been evaluated in the context of cell patterning, a method to reconstruct structure of biological tissue, cell by cell. There, using very high intensity laser pulses, individual cells are flung forward utilizing the impulse force generated by focused femtosecond laser irradiation. It has been shown that the biological activity of the cell was not affected by laser patterning [13]. On the normative point of view, the maximum electromagnetic radiation to which human tissues can be exposed has been estimated as a function of the electromagnetic wavelength, exposure time, and pulse repetition. The International Electrotechnical Commission [14] and the American National Standard [15] define maximum permissible exposure (MPE) levels. For instance, for a 400 nm wavelength the stated MPE for pulse duration of 1– 100 ns is 20 mJ/cm2. The fluence we are using in our experimental set up is by three order of magnitude less (15 lJ/cm2) since the energy per pulse is in the pJ range and the beam is focused down to an area of few square micrometer.

4.2. Temperature rise Fig. 4 shows the temperature rise caused by a laser pulse, at different times, at the interface vicinity in the substrate and the cell. For these calculations, water was considered instead of cell since these media have close mechanical and thermal properties. The parameters for water and titanium used for the calculations are given in Table 1. No thermal resistance was considered at the interface, thus leading to an overestimation of the temperature rises. Fig. 4 shows that the maximum temperature rise is 1 K only during the very first picoseconds and that it cools down to 0.3 K in less than 100 ps. Moreover, the heated depth in the cell is one tenth of that in the metal, about 10 nm only. The question arises of the steady state temperature rise due to the repetitive heating of the surface. This heating is estimated considering a CW laser source (of a power 50 lW) with a modulation frequency of fm = 300 kHz. A 3D modelling of the temperature rise on the surface of an isotropic half space by a modulated laser with a Gaussian profile is considered. Within a low frequency approximation [16], it gives a temperature rise of 0.5 K. This temperature rise is overestimated since an air–titanium interface was assumed, neglecting thermal diffusion in the transparent medium.

1

0.8

0.6

cell

metal

(K) 0.4

0 -50 -100

(Pa) -150 -200 -250 -300 0

2

4

6

8

10

12

14

(ps) Fig. 5. Normal stress at the interface for thermal resistance R = [1, 5, 10] 109 K W1. The stress magnitude deceases when R increases.

The temperature rise in water can be measured experimentally since the Brillouin frequency, i.e. the sound velocity in water is related to temperature. A temperature rise of 1 ± 1 K was measured in the same experimental conditions. 4.3. Stress The total stress ri(z, t) was calculated at the interface position (z = 0) for several values of the thermal resistance R. The negative stress, Fig. 5, indicates a force oriented toward the cell. The maximum stress peak magnitude is about 200 Pa and the stress pulse duration is extremely short, typically a few picoseconds. For instance, adhesion forces measured for osteoblast cells on glass [17] are about 1 lN, thus giving adhesion stresses of several kPa. The applied stress calculated at the interface is thus lower than this adhesion stress, itself being lower than the cell adhesion stress on a bio-material such as titanium. 5. Reproducibility and sensitivity To illustrate the reproducibility and the sensitivity of the method, experiments have been achieved on two sets of yellow and white allium cepa cells, each set including 10 cells extracted from 10 allium cepa bulbs. The measured frequencies and phases of the Brillouin oscillations measured in cells vacuole and cells nucleus are shown in Fig. 6. Owing to the similarities of the mechanical properties, results measured in the vacuoles of the two allium cepa cell varieties are equal. Frequencies measured in the nuclei differ sensitively from that obtained in the vacuoles. This point will be commented in the following section. In addition, the Brillouin frequencies measured in the nucleus of either yellow or white allium cepa cells are significantly different. Therefore, the values of these frequencies allow to differentiate two varieties of the same cell species from the measured acoustic signature in an organelle. 6. Toward single cell imaging

0.2

0 -40

50

0

40

80

120

160

200

(nm) Fig. 4. Temperature rise in the vicinity of the cell–substrate interface from time t = 0–100 ps with a 20 ps step.

A 25 lm, 1D scan in a white onion is performed with five lateral steps of 5 lm each to perform six measurements with a lateral resolution of 2 lm. The first measurement is made in the nucleus of the cell, the five others through the vacuole. Four acoustic parameters of the Brillouin oscillation are recorded and plotted in Fig. 7, i.e. frequency, phase, amplitude and attenuation. The acoustic parameters are very different in one hand for the nucleus and in the other hand for the vacuole over 20 lm.

206

B. Audoin et al. / Ultrasonics 50 (2010) 202–207

Nucleus 4

4

3

3

3

3

2

2

0 5.2

0 5.6

6

6.4

6.8

7.2

0 5.2

/2

0 5.6

phase

6

6.4

6.8

7.2

/2

frequency (GHz)

phase

4

4

3

3

3

3

2

2 1

1

5.2

5.6

6

6.4

6.8

2

0

0

7.2

5.2

/2

frequency (GHz)

2 1

1

0

0

Count

4

Count

4

Count

Count

2 1

1

frequency (GHz)

White

2

1

1

Count

4

Count

4

Count

Count

Yellow

Vacuole

5.6

phase

6

6.4

6.8

7.2

/2

frequency (GHz)

phase

7

200

6.5

150

phase (°)

frequency (GHz)

Fig. 6. Frequencies and phases of the Brillouin oscillations measured in cells vacuole and cells nucleus for two sets, each including 10 cells, of yellow and white allium cepa cells.

6

5.5

100

50

5 0

5

10

15

20

25

0

5

position ( m)

10

15

20

25

20

25

position ( m) 2

15

attenuation (GHz)

-7

amplitude (10 )

1.5

10

5

1

0.5

0

0 0

5

10

15

position ( m)

20

25

0

5

10

15

position ( m)

Fig. 7. Scan inside a single onion cell. Frequency, phase, amplitude and attenuation are obtained from the Brillouin oscillations.

B. Audoin et al. / Ultrasonics 50 (2010) 202–207

Brillouin frequency oscillations is related, see Eq. (7), to the optical index number and to the acoustic celerity. For the nucleus and for the vacuole the mean value of the Brillouin frequency is equal to 6.8 ± 0.2 GHz and 5.7 ± 0.2 GHz, respectively. Taking a typical optical index close to that typical for cells, as n = 1.4, the average velocity of ultrasound in vacuole and in the nucleus is obtained as 1.6 ± 0.1 lm/ns and 1.9 ± 0.1 lm/ns, respectively. Interpretation of the phase of the Brillouin oscillation is not straightforward since it depends on several physical parameters including optic and acoustic attenuations. The phase of the Brillouin oscillation is measured equal to 70° and 172° for the nucleus and the vacuole, respectively. The difference may be caused by different optical absorptions in the nucleus and the vacuole. The amplitude of the Brillouin signal in the nucleus is significantly lower than the amplitude of the signal in the vacuole because of two main reasons: the nucleus is semitransparent and it is not directly in contact with the substrate. The acoustic attenuation limits, for a given acoustic amplitude, the depth of the cell which could be probed by the picosecond ultrasonics technique. Indeed, the Brillouin oscillations end at 1.2 ns corresponding to 2 lm depth using the acoustic velocity previously measured. For comparison, water has an acoustic attenuation value which permits to probe more than 5 lm in-depth [3]. The measured acoustic attenuation is 0.1 ± 0.1 GHz and 1.5 ± 0.1 GHz for the vacuole and the nucleus, respectively. Fig. 7 shows that the measured frequencies and phases of the Brillouin oscillations are very similar for all the scanned part of the vacuole (20 lm): the acoustic velocities and the optical index are also very similar in all the vacuole. However, this figure shows that the measured amplitudes and attenuations of the Brillouin oscillations have a larger dispersion range in the scanned vacuole. This dispersion can be related to the quality of the local mechanical adhesion of the cell on the substrate or the surroundings. 7. Summary and perspectives Very promising abilities of the picosecond ultrasonic technique for the non-invasive study of biological alive single cells have been demonstrated in this paper. The measurements involving lasers for both the opto-acoustic generation and the acousto-optic detection avoid technological limitations in frequency encountered in acoustic microscopy when piezo transducers are used for both generation and detection, as well as when laser pulses are used for wave generation only. Acoustic frequencies a high as 11 GHz are reported in this paper for the first time in a biological medium, i.e. a vegetal cell. A lateral resolution of 1 lm, and a depth resolution better than 100 nm can be achieved with this technique. Several perspectives can be foreseen for the picosecond biophononic technique. In the context of emerging nanomedicine science and technology, experimental determination of physical properties of alive single cells is an important step toward better understanding of cell properties. Knowledge of mechanical properties such as cytoplasm compressibility and effective cell viscosity can help to understand and model cellular and intracellular processes. The application of contactless and very high frequency opto-acoustic technique will improve significantly the space resolution of acoustic imaging of alive cells. It will also provide more information

207

about mechanical properties of the cell. From the measured velocities, cell compressibility nearby the cell–substrate interface can now be mapped with a resolution better than 1 lm laterally and 100 nm in-depth. In addition to single cell elastography, the sensitivity of the measurements to cell adhesion [5] suggests promising perspectives for the imaging of the physiological functions of the cell, in function of its surroundings or of its health, with the aim of a better understanding of the mechanism of the cell adhesion to the implant. Other applications could be considered in medicine. Some diseases, such as cancer, provoke important decrease of the cell compressibility. All the same, Alzheimer’s disease is associated with a change in the stiffness of the neuronal cell cytoskeleton. The acoustic elastography: (i) could help diagnosis and (ii) could help analyzing the effect of the introduction of active substances. The dynamics of the heterogeneous effect of these substances inside a single cell could be mapped via measurements of the acoustic parameters within the cell. References [1] C. Thomsen, J. Strait, Z. Vardeny, H.J. Maris, J. Tauc, J.J. Hauser, Coherent phonon generation and detection by picosecond light pulses, Phys. Rev. Lett. 53 (1984) 989. [2] B. Perrin, C. Rossignol, B. Bonello, J.-C. Jeannet, Interferometric detection in picosecond ultrasonics, Physica B 263–264 (1999) 571. [3] L.J. Shelton, F. Yang, W.K. Ford, H.J. Maris, Picosecond ultrasonic measurement of the velocity of phonons in water, Phys. Stat. Sol. B 242 (2005) 1379. [4] O.B. Wright, B. Perrin, O. Matsuda, V.E. Gusev, Optical excitation and detection of picosecond acoustic pulses in liquid mercury, Phys. Rev. B 78 (2) (2008) 024303. [5] C. Rossignol, N. Chigarev, M. Ducousso, B. Audoin, G. Forget, F. Guillemot, M.-C. Durrieu, In vitro picosecond ultrasonics in a single cell, Appl. Phys. Lett. 93 (2008) 123901.1–123901.3. [6] C. Thomsen, H.T. Grahn, H.J. Maris, J. Tauc, Surface generation and detection of phonons by picosecond light pulses, Phys. Rev. B 34 (6) (1986) 4129. [7] J. Beuthan, O. Minet, J. Helfmann, M. Herrig, G. Muller, The spatial variation of the refractive index in biological cells, Phys. Med. Biol. 41 (1996) 369. [8] C. Saubade, Refringency law and optical properties of water at various temperatures: II. Macroscopic approach, J. Phys. C: Solid State Phys. 17 (1984) 3507. [9] W. Denk, J.H. Strickler, W.W. Webb, Two-photon laser scanning fluorescence microscopy, Science 248 (1990) 73–76. [10] D.W. Piston, M.S. Kirby, H. Cheng, W.J. Lederer, W.W. Webb, Two-photon excitation of fluorescence imaging of three-dimensional calcium-ion activity, Appl. Opt. 33 (1994) 662–669. [11] E. Berry, G.C. Walker, A.J. Fitzgerald, N.N. Zinov’ev, M. Chamberlain, S.W. Smye, R.E. Miles, M.A. Smith, Do in vivo terahertz imaging systems comply with safety guidelines?, J Laser Appl. 15 (3) (2003) 192–198. [12] K. König, P.T.C. So, W.W. Mantulin, E. Gratton, Cellular response to nearinfrared femtosecond laser pulses in two-photon microscopes, Opt. Lett. 22 (2) (1997) 135–136. [13] N. Takizawa, K. Okano, T. Uwada, Y. Hosokawa, H. Masuhara, Viability evaluation of culture cells patterned by femtosecond laser-induced impulsive force, Progress in Biomedical Optics and Imaging – Proceedings of SPIE 6854 (2008) 685411. [14] International Electrotechnical Commission, International Standard, Safety of Laser Products – Part 1: Equipment Classification and Requirements, IEC 60825-1, ISBN 2-8318-9085-3. [15] American National Standards Institute, American National Standard for the Safe Use of Lasers in Health Care Facilities: Standard Z136.1, ANSI Inc., New York, 2000. [16] D.G. Cahill, Analysis of heat flow in layered structures for time-domain thermoreflectance, Rev. Sci. Instrum. 75 (12) (2004) 5119. [17] D. Debavelaere-Callens, L. Peyre, P. Campistron, H.F. Hildebrand, On the use of ultrasounds to quantify the longitudinal threshold force to detach osteoblastic cells from a conditioned glass substrate, Biomol. Eng. 24 (5) (2007) 521–525.