# Distributed 2D temperature sensing during nanoparticles assisted laser ablation by means of high-scattering fiber sensors

Jul 28, 2020

### Experimental setup

The experimental set up of multiplexed temperature monitoring during laser ablation of the porcine liver phantom using nanoparticles gold and magnetic iron oxide is represented schematically in Fig. 1, with the photographic view in Fig. 2. The sensing design consists of: (1) an Optical backscattering reflectometry (OBR), (2) fiber-coupled laser diode, (3) data processing software, (4) four nanoparticle doped optical fibers (NPDF), (5) fresh porcine liver.

The ablation procedure was conducted by means of fiber-coupled mid-power laser operating in a continuous mode (980 nm, Roithner Lasertechnik GmbH, Austria). The laser was calibrated and set at a current of 2.34 A which corresponding to 4 W and the ablation procedure lasted from 120 to 145 s. The laser light was delivered to the liver phantom through 1 m long fiber applicator with the core diameter 400 µm and numerical aperture 0.22 at a 75 mm distance in the z-direction (Fig. 2). The delivery fiber is clearly multimode and emits with a very large divergence angle θ. The beam waist can be approximated to the radius of the fiber 200 µm. The divergence angle θ is related to the aperture number NA with the well know relation: sinθ = NA. The spot size radius can then be estimated as:

$$0.2;{text{mm}} + 75;{text{mm}} times tanuptheta = 0.2;{text{mm}} + 75,{text{mm}} times {text{NA/}}surd left( {1 + {text{ NA}}^{2} } right) = 17.1;{text{mm}}$$

This is an approximated assumption, however, it is similar to the laser ablation setup with frontal firing fibers (as in the example Table 3 of ref.3).

The laser applicator was fixed using a holder on a perpendicular direction to the tissue as shown in Fig. 2 to conduct the laser ablation on the tissue surface. The thermal ablation process was done following the European protocol of “Three Rs”29 on a commercially available porcine liver purchased from the butchery shop and maintained at a room temperature 22–25 °C right prior to the experiments. The liver phantom stabilized until room temperature for several hours in order to mimic a real application close to the body temperature. The initial temperature was monitored using contact thermocouple (IKA ETS – D5). Moreover, a FLIR thermal camera was used to validate that the tissue temperature distribution is uniform, at least within the accuracy of the thermal camera. The accuracy of all FLIR cameras for IR detection at temperatures 0–120 °C ranges from ± 2 to ± 3 °C.

Temperature measurements are performed with distributed fiber optic-based sensing setup, consisting of commercial OBR (OBR4600, Luna Inc.) to interrogate spectra of the fiber parallel. The fiber parallel is made up of 4 SMF-28 extenders each spliced to a piece of NP-doped fiber using the Fujikura 12-S fusion splicer. A 1 × 8 splitter is used to create 4 lines of the parallel in the Y direction. Distributed temperature measurements taken at longitudinal points on the fibers were used to fill the 2D-map as shown in Fig. 3.

In this setup, temperature changes are measured by analyzing the backscattering profiles of NP-doped fibers, while SMF serves as connecting elements. The backscattering profile is shifted in response to a temperature change so that it can be determined by comparing the measured spectra with the initial, or reference trace. It is important, that backscattering profiles of each of 4 NP-doped fibers do not overlap; to prevent such overlapping the length of each SMF-28 separator must be longer than the previous.

OBR based on the principles of Optical Frequency Domain Reflectometry, which allows employing single-mode fiber as a valuable spatially distributed temperature sensor24. OBR has a great advantage for the thermal ablation application due to its high resolution (less than 1 cm)21. However, a serious limitation of OBR is its inability to provide multiple sensing fibers. Thermal ablation requires spatial two-dimensional temperature measurement, which cannot be achieved by a single fiber. Therefore, in this experimental set up we used nanoparticles-doped fibers (NPDF), which have about 30 dB higher scattering level than the standard single-mode fibers (SMF-28) as shown in Fig. 4. The NPDF fiber has been fabricated in two steps: at first, using modified chemical vapor deposition (MCVD) a silica preform was fabricated, with the addition of MgCl2 and ErCl3 dopants; then, the fiber was drawn in a standard drawing tower. The high temperature up to 2000 °C causes a separation of silica and alkaline ions, forming MgO-based nanoparticles with 20–100 nm diameter that elongate along with the fiber through the drawing process30,31. Four NPDFs with the core diameter of 10 µm and 125 cladding diameter are spliced with SMF pigtails of different lengths so that two neighboring NPDFs are spatially separated by the SMF fiber from the neighboring NPDF. The presence of MgO NP induced a much larger Rayleigh scattering. The spectrum, illustrated in Fig. 4, shows that the desired multiplexing is achieved because NPDFs are separated by the regions of lower scattering power of SMF28.

The 4 NPDF sensing fibers have been positioned on the liver porcine tissue in situ at a distance 0.5 cm that can be seen from photographic view in Fig. 2 from each other in y-direction. Such positioning allows monitoring the changes in temperature not only at the center of the ablation area but also in the peripheral area. Figure 3 shows graphically the positions of the fibers with respect to the tissue: the lines report the 4 fibers, and the dots the sensing points along the fibers detected by the OBR. The temperature sensing technique consists of 52 sensing points over the 5.4 cm2 area.

All trials were performed at the fixed settings of laser and optical fiber sensing parameters with the in situ positioning of fibers on the tissue. However, the only changing condition during the LA was the treatment of the tissue with nanoparticles and without nanoparticles. In addition, two different types of metallic nanoparticles were employed, namely gold and magnetic iron oxide nanoparticles. Both nanoparticles were spherical in shape and had a size of around 20 nm. Magnetic iron oxide nanoparticle was diluted in a 0.2% agarose solution with a concentration of 5 mg/ml. The aqueous stock solution of gold nanoparticles with a concentration of 0.055 mg/ml was used in the ablation procedure. The 100 µl of nanoparticle solution (both gold and magnetite nanoparticles) from previously prepared stock were injected ex vivo onto the liver phantom using the pipette. This setup, having a short distance between the fiber output and the tissue, has been designed to mimic noninvasive and superficial laser ablation11,13.

### Data acquisition and analysis

Temperature values were collected along the lengths of the fibers by the optical backscatter reflectometer operated in a distributed sensing mode. The step size between each temperature measurement was 0.3 cm. Spaces between each pair of the fibers were 0.5 cm, and the length of each fiber segment under measurement was 3.6 cm. The alignment of the fibers is schematically represented in Fig. 3.

Data acquisition was conducted at a rate of one measurement/s. Then, the resulting temperature measurements were interpolated with the spline method, introducing additional data points at intervals of 0.02 cm across the XY-plane in Fig. 3. The ‘pixel size’ is 3 mm (X) × 5 mm (Y), so 15 mm2 areas over a surface of 36 mm ×15 mm = 540 mm2. The 2D fit interpolates the data on a grid of 0.2 mm × 0.2 mm = 0.04 mm2 (13,500 interpolated sensing points).

Each value of temperature was then assigned a specific color. The resulting 2D colored maps were saved for all points in time throughout each experiment on ablation. In addition, quiver plots were produced for all of the thermal maps to clearly indicate the direction of temperature change in the experiments.

### Fiber calibration

Four NPDF fibers used in the experiments were thermally calibrated first. They were immersed in a flask filled with water. A fiber containing one FBG (Technica S.A., 10.2 pm/°C) was also placed there to provide a reference. The water inside the flask was heated while monitoring the spectra recorded by the 4 fibers and the FBG. Temperature changes were calculated using the wavelength shifts of the FBG. Then, temperature changes indicated by the FBG were compared with the shifts recorded by four fibers. The results, including the sensitivities of all fibers, can be seen in Fig. 5. According to the calibration results, all the fibers have the same sensitivity coefficient, which is equal to 10.42 pm/°C and is similar to FBGs written on standard glass fibers.

### Synthesis and characterization of nanoparticles

#### Gold nanoparticles

The synthesis of gold nanoparticles was conducted in a three-necked 250 mL round-bottomed flask cleaned with a dichromate solution according to the method of32. The set up was placed on a hot plate fitted with a magnetic bar for continuous stirring in the silicon oil heating condition. 0.5 ml of 1% stock solution of hydrogen tetrachloroaurate (III) trihydrate first added into the flask through funnel followed by 50 ml of distilled water (DI). The hot plate adjusted to the temperature of 105–108 °C. The resulting solution was heated until boiling. Afterward, 2.0 ml of 34 mM citrate solution was added to the solution. The solution was refluxed further for 15 min under stirring conditions and slowly cooled down to room temperature under stirring. Finally, the obtained gold nanoparticles were washed two times with DI water using a centrifuge for 15 min under 12,000 rpm and stored in vials at room temperature. The color of the final nanoparticle solution was ruby-red. UV–VIS spectral analysis and TEM scanning were performed for each set of experiments. All chemicals were of analytical grade and purchased from the Merk Company.

For TEM analysis the samples were prepared by dropping the aqueous solution of gold nanoparticle solution on 400 copper mesh and left for drying at room temperature. The samples were analyzed on Transmission Electron Microscope (JEOL JEM—1400 Plus). The obtained TEM micrographs of gold nanoparticles demonstrate the homogenous distribution of the nanoparticles with the size range between 14 and 20 nm. The shape of the gold nanoparticles is spherical according to the obtained data in Fig. 6a.

Figure 6b shows the absorption spectra of synthesized gold nanoparticles analyzed on Evolution 300 UV–Vis Spectrophotometer. The absorption maximum of AuNPs was 520 nm, which stands for nanoparticles of size 20 nm33. This is in agreement with the expected ruby-red color of the solution, which also indicates that the size of nanoparticles is 20 nm.

#### Magnetic iron oxide magnetic nanoparticles

The iron oxide magnetic nanoparticles were synthesized by the solvothermal method adapted from the method34. 2.535 g of FeCl3*6H2O and 1.8625 g of FeCl2*4H2O were dissolved in 6.25 ml of distilled water under the magnetic stirring in a beaker followed by the addition of 6.25 ml of 25% Ammonium Hydroxide. The obtained solution was stirred at 700 rpm for 2 min. Then placed into the Teflon-lined stainless steel autoclave and heated for 1 h at 180 °C in a muffle furnace. The synthesized nanoparticles were washed several times with distilled water and dried for further characterization and application for the thermal ablation. The X-Ray Diffraction (XRD) and Scanning Electron Microscopy (SEM) analysis were done in order to validate the size and shape of nanoparticles. All chemicals were purchased from Merk Company. The use of TEM analysis was avoided for magnetite due to its magnetic property.

The crystal structure and the size of the synthesized Fe3O4 nanoparticles were analyzed by the Rigaku SmartLab X-ray diffraction (XRD) system in Fig. 7a. The diffraction peaks obtained at 30.1°, 35.4°, 43°, 53.4°, 56.9°, 62.5° meet the diffractions values of [220], [311], [400], [422], [511] and [440] planes of Fe3O4 crystals respectively. The obtained XRD patterns matched the literature value for magnetite peaks35,36.

The size of the iron oxide magnetite was calculated as 20 nm using Scherrer’s equation. While, the surface morphology of iron oxide nanoparticles was observed using Scanning Electron Microscope (SEM, Auriga Crossbeam 540), which shows mostly uniform distribution with the size around 20–40 nm as demonstrated in Fig. 7b.