The design of the quantitative phase laser microsurgery system is described in . After passing through a 5X beam expander, the second harmonic 532 nm green scissor beam from a nanosecond Nd: YVO4
laser (20KHz, 12 ns, Coherent Inc, USA) is guided through a shutter and a lens pair, deviated by a mirror (M3) and focused onto a small spot in the sample by an objective (40X, 0.65NA). The lens pair, composed of a concave and a convex lens, is introduced to match the focal plane with the imaging plane of the objective. For the purpose of quantitative phase imaging, the red light from a laser diode of 675 nm is focused by a lens, L, and split into a Mach-Zehnder interferometer. The two light beams through the object and the reference arms are directed towards a CCD camera (Sony XCD 710). A 20X objective was placed in the reference arm of the interferometer and the curvature of the interference fringes was digitally compensated and controlled via software [24
]. A slight angle is introduced between the object and the reference beams for off-axis holography. The camera has an array of 1024 × 768 pixels with pixel size of 4.65 μm and 8-bit gray scale output with an acquisition speed of 30 frames per second. An IEEE1394 cable connects the camera to the desktop computer, which processes the acquired images and calculates the holographic diffraction using a number of programs based on LabVIEW® in real-time. From Fourier optics [25
], if E
;0) is the object wave field at plane z
= 0 , the corresponding angular spectrum of the object wave at this plane, S
;0), can be obtained by taking the Fourier transform of the hologram and separating it from other spectral components of the hologram with a band-pass filter if the off-axis angle θ
of the incident beam is properly adjusted. Here kx
are corresponding spatial frequencies of x
. After propagating along the z
axis to a new plane, the new angular spectrum, S
), at plane z
can be calculated from S
;0) as S
], where kz
. Thus the complex field distribution of any plane z
perpendicular to the propagating axis can be calculated by an inverse Fourier transform of S
). The resolution of the reconstructed images from the angular spectrum method is the same as that in the hologram plane. The nonambiguity phase range calculated from the complex field distribution is only from −π
. Any phase outside this range will cause a wrapping effect of the phase map. The 2π
-phase ambiguities can be directly resolved to get an absolute sample phase map by phase unwrapping using Goldstein’s algorithm [26
Schematic of the experimental setup. LD is the 675nm laser diode; BS1 and BS2 are beam splitters; Ms are mirrors; S is a shutter; L is a lens; HPF is a high-pass filter for red light.
We present several examples of the quantitative phase laser microsurgery system for evaluation of cellular/ sub-cellular dynamic changes during laser micro-dissection. shows images of rat kangaroo kidney epithelial (PTK2) cells ablated with a laser microbeam. The bright-field image of the cell before the laser microbeam is shown in , which shows many chromosomes lying inside the cell. The laser beam was targeted at one spot of the chromosomes as marked with a white arrow in . A custom-designed program was used to record holograms in real time and could be triggered to record automatically all of the phase maps of the PTK2 sample before and after laser microsurgery. With a pulse frequency of 20 kHz, the average power of the laser microbeam coming out of the 40X objective was measured to be 8 mW, which corresponds to 0.4 μJ/ pulse. With a pulse width of 12 ns, this corresponds to a peak irradiance of 9 2 2.67×109W / cm2. The bright-field image right after the laser microbeam is shown in , from which one can observe the defocusing of the chromosomes which could be because of lifting up of the chromosomes. This defocusing may be associated with swelling of the cell. However, from the bright field images, the swelling cannot be confirmed. After about 5 minutes, the defocusing of the chromosomes vanished as shown in . shows several time-lapse quantitative phase maps of the PTK2 cell during laser microirradiation, assuming the shutter open time as the zero time point. also shows a movie of quantitative holographic phase maps of the cell during and after laser microirradiation. The relative phase changes of the cell compared to its original status before laser microirradiation is also given in . From the quantitative phase and phase change information, the change in physical thickness of cells can be estimated using the equation Δd = λ(Δϕ/2π) / (n − n0), where λ is the wavelength, Δϕ is the phase, and (n − n0) is the index difference between the cells and the buffer (or other media). The phase map in ranges from 0 to 3 radians, as indicated in the color bar, which corresponds to an optical thickness span of 6.4 μm, with the assumption of the average index of the cell to be 1.38 and that of the buffer to be 1.33. Though the mechanism for the swelling (by heating and by expansion of the cytoplasm) is still under investigation, the laser-induced swelling can now be quantitatively evaluated. As shown in , the maximum phase change of ~1.0 radian corresponds to an optical thickness change of 2.1 μm assuming that the refractive index of the cell near the laser microirradiated region does not change. shows the laser micro-dissection of the nucleolus of a PTK2 cell. The laser microbeam is scanned to cut a portion of the nucleolus by manually scanning the mirror (M3). When the laser microbeam is on, we allowed (through the high pass filter) a small portion of the green microbeam to reach the camera so as to enable visualization of the focused spots (shown as bright spots in ). If the green microbeam is bright enough to nearly saturate the focused spots in the interferogram, they appear as “singular” (non-smooth) phase points in the phase map, clearly indicating the ablation/focused laser position, as shown in . Note that the ablated nucleolus in may have a different refractive index from other parts of the cell. To be more accurate, only the phase scale (in radians) is given in . The actual ablation/profile (in microns) can be estimated similarly using a proper estimation of the refractive indices for the cutting spot based on the obtained phase information.
Fig. 2 (Color) Bright-field images of rat kangaroo kidney epithelial (PTK2) cells: (a) Before the laser microbeam; (b) 30 seconds after the laser microbeam; (c) 5 min after the laser microbeam. Figure (d) shows several time-lapse quantitative phase images during (more ...)
Fig. 3 Bright-field images of PTK2 cell nucleolus: (a) before laser micro-dissection; (b) during laser micro-dissection, with the targeting laser beam on (and passed to the camera) showing the micro-focused spot; (c) after laser micro-irradiation. Figure (d) (more ...)
shows an example of dynamic microsurgery of red blood cells harvested from a healthy volunteer. The RBCs were separated by centrifugation at 3000 rpm for 10 min and suspended in Phosphate Buffered Saline (pH = 7.4, n = 1.33). The RBC suspension was diluted in phosphate buffer (350 mOsm/Kg), spread on a coverslip and observed by the QPLM system. The bright-field image of the crenate RBCs is shown in . The position of the laser microbeam in this case is controlled electronically by the mirror (M3) and scanned to cut a straight line across a single RBC. The whole process takes about 4 seconds. shows several time-lapse quantitative phase maps of the targeted RBC ablated by the scanning laser microbeam, assuming the ablation starting point as the zero time point. shows a movie of quantitative phase changes of the ablated cell.
Fig. 4 (Color) (a) Bright-field image of crenate RBC cells; and (b) time-lapse quantitative phase maps of the RBCs during laser microirradiation. Figure (c) shows a movie of quantitative phase images before and during laser microbeam (Media 4). Assuming n = (more ...)
Finally in , we show quantitative phase imaging of the response of Goldfish retinal rod cells to green light (laser) exposure. It may be noted that the rod cells in the retina contain photoreceptors that are very sensitive to light, and hence over-exposure of light (as exists in laser microbeams) is expected to cause irreversible damage to the rod cells. Since phase imaging is very sensitive to very subtle changes in the microscopic domain, we should be able to determine the threshold of light required for damaging the rods. For this purpose, we used a low power laser microbeam (0.1 μJ/ pulse) to hit one end of the rod (as indicated in ). The cell membrane of the end was slightly damaged. The rod responded immediately to the irradiation leading to drastic morphological changes. show the brightfield images of the rod cell before and after laser irradiation. A time-lapse movie in shows the response of the rod to the laser irradiation and associated morphological changes. Note that our samples are prepared in plastic round chambers with open fluid (buffer). The vibration of the fluid surface will bring some residual phase noise, as is evident in some of the above phase change movies. The system performance could be improved by using closed cell chambers and controlling mechanical vibrations.
(Color) Bright-field images of the rod cell from Goldfish retina: (a) Before the laser microbeam; (b) after the laser microbeam; Panel (c) shows quantitative phase change during laser microirradiation (Media 5). Scale bar: 10 μm, n = 1.38.