Plasma Science and Technology, Vol.18, No.6, Jun. 2016

Investigation of Laser Induced Breakdown Spectroscopy (LIBS) for the Differentiation of Nerve and Gland Tissue–A Possible Application for a Laser Surgery Feedback Control Mechanism 2 ¨ F. MEHARI1,2 , M. ROHDE1,3 , C. KNIPFER3 , R. KANAWADE1,2 , F. KLAMPFL , 4 3 1,3 1,2 W. ADLER , N. OETTER , F. STELZLE , M. SCHMIDT 1

Clinical Photonics Lab, Erlangen Graduate School in Advanced Optical Technologies (SAOT), Friedrich-Alexander-Universit¨ at Erlangen-N¨ urnberg, Paul-Gordan-Str.6, 91052 Erlangen, Germany 2 Institute of Photonic Technologies, Friedrich-Alexander-Universit¨ at Erlangen-N¨ urnberg, Konrad-Zuse-Str.3/5, 91052 Erlangen, Germany 3 Department of Oral and Maxillofacial Surgery, University Hospital Erlangen, FriedrichAlexander-Universit¨ at Erlangen-N¨ urnberg, Gl¨ uckstrasse 11, 91054 Erlangen, Germany 4 Chair of Biometry and Epidemiology, Friedrich-Alexander-Universit¨ at Erlangen-N¨ urnberg, Waldstraße 6, 91054 Erlangen, Germany

Abstract

Laser surgery provides clean, fast and accurate modeling of tissue. However, the inability to determine what kind of tissue is being ablated at the bottom of the cut may lead to the iatrogenic damage of structures that were meant to be preserved. In this context, nerve preservation is one of the key challenges in any surgical procedure. One example is the treatment of parotid gland pathologies, where the facial nerve (N. VII) and its main branches run through and fan out inside the glands parenchyma. A feedback system that automatically stops the ablation to prevent nerve-tissue damage could greatly increase the applicability and safety of surgical laser systems. In the present study, Laser Induced Breakdown Spectroscopy (LIBS) is used to differentiate between nerve and gland tissue of an ex-viv o pig animal model. The LIBS results obtained in this preliminary experiment suggest that the measured spectra, containing atomic and molecular emissions, can be used to differentiate between the two tissue types. The measurements and differentiation were performed in open air and under normal stray light conditions.

Keywords: laser induced breakdown spectroscopy, plasma, chemical composition, nerveparotid gland

PACS: 32.30.Jc, 87.64.−t, 52.25.Os DOI: 10.1088/1009-0630/18/6/12 (Some figures may appear in colour only in the online journal)

1

Introduction

nomenclature, several different surgical approaches are described in the literature, ranging from a superficial parotidectomy with the dissection and preservation of the facial nerve (N. VII) to a radical total parotidectomy with sacrifice of N. VII, depending on different factors such as the location and/or type of the tumor [7−10] . Both, recrudescence due to incomplete tumor removal [11] and facial nerve damage [12] can pose a threat to the patient’s wellbeing and must be avoided if possible.

Pathologies of the salivary glands range from bacterial/viral infections, sialadenitis and congenital diseases to benignant and/or malignant tumors [1,2] . Benignant tumors mostly occur in the parotid gland and make up a considerable portion of all head and neck neoplasms [2] , whereas malignant tumors are reported to be less common [3,4] . The serous Glandula parotidea is the largest of the salivary glands [1] and shows a close anatomic relationship to the facial nerve, which exits the scull through the stylomastoid foramen and divides into its main branches within the gland to innervate the musculature of the face [5,6] . The main goal of any surgery performed on the parotid gland is the complete removal of the affected tissue without compromising facial nerve function [7] . Here, in an“increasingly complicated” [7]

In this context, laser based surgery is an interesting tool, as it offers a number of advantages when compared to conventional surgery: surgical laser systems allow the ablation of very thin layers of tissue [13] while the induced coagulation of blood vessels permits a dry operating field with better visibility [14−18] . At the same time, lymphatic vessels and salivary ducts are sealed 654

F. MEHARI et al.: Investigation of LIBS for the Differentiation of Nerve and Gland Tissue and may prevent the formation of postoperative sialoceles [19] . Soft tissue manipulation using laser systems on oral tissues was also shown to result in good healing as well as less postoperative swelling and inflammation when compared to conventional surgery [20−24] . Contact based Nd:YAG laser parotidectomies have been reported [19] . However, both contact and contactless surgical laser systems face the inability to determine the type of tissue that is being ablated at the bottom of the cut and suffer from a lack of feedback for the surgeon. In a noncontact surgical laser, the actual penetration depth is unknown and may lead to iatrogenic damage of nerves, blood vessels or other important anatomical structures [25−28] . To overcome this problem, the authors have previously proposed a laser induced breakdown spectroscopy (LIBS) based feedback control mechanism, directly using the process emissions for the differentiation of soft and hard tissues [29,30] . To use LIBS for the differentiation of tissues, a short pulsed laser is focused on the sampl’s surface. Once the samples are exposed to a pulse of sufficient energy, ablation of the material can be observed. This is followed by the formation of plasma as a result of the matrix effects initiated due to the further absorption of the laser light by the ablated material [31] . The generated plasma then, through radiative and other processes, one of which is atomic emission, relaxes in the ambient air [32] . They result from the electronic transitions between different elemental excitation states. Spectrometric measurement targeted at these atomic emissions can provide qualitative and relative quantitative information on the elemental composition of the investigated sample. The relative quantitative information is based on the higher atomic-emission-intensity readings of those elements with a relatively higher concentration [33,34] . In addition, properly timed measurement can also provide emissions from the molecular species in the plasma [32] . The aim of the present study is to differentiate between nerve and gland tissue by monitoring the plasma generated during a nanosecond pulse laser-mediated ablation of the tissues. A successful real time differentiation of the tissues may help broaden the scope of applications and increase the safety of clinical laser systems.

2 2.1

1 cm was chosen as it was easy to identify and access in the test animals. After incision in the upper mucobuccal fold, the soft tissue was lifted off the maxillary bone using a raspatory, and the Nervus infraorbitalis was dissected using anatomical scissors. The resulting samples had an average length of 6 cm and a diameter of approximately 1 cm. For the gland tissue, the parotid gland of the animals, located caudal and buccal of the ascending branch of the mandible was carefully freed from the surrounding soft tissue. After sagittally splitting the mobilized glands, samples of approximately 2×2 cm2 were prepared. After preparation, the samples were rinsed with sterile 0.9% saline solution to remove any remaining superficial contaminants and then stored in an opaque, air tight container at 4 o C.

2.2

Experimental setup

Fig. 1 shows the schematics of the experimental LIBS setup. A frequency doubled Nd:YAG laser (Saga Flashlamp pumped Nd:YAG, Thales Group, Neuilly-SurSeine, France), working at 532 nm with the pulse frequency set to 1 Hz and a pulse duration of 10 ns was used in all of the measurements. The average pulse energy was determined to be 80 mJ. To focus the laser beam on the tissue samples (approximately 0.3 mm focus size), a 50 mm focal length lens was inserted in the optical path and a transitional stage was used to bring the sample’s surface to the focal spot.

Fig.1

Materials and methods Animal model and tissue sample preparation

Schematics of the LIBS experimental setup

During the plasma generation, the resulting plasmalight of each of the pulses was collected using a 50 µm fiberoptic cable connected to a spectrometer (Mechelle Me 5000 Echelle, Andor, Belfast, UK). The spectrometer is equipped with an ICCD camera (A-DH334T18F-03 USB iStar ICCD detector, Andor, Belfast, UK) and has a spectral resolving power (λ/∆λ) of 6000 in the range of 200-975 nm. To synchronize the measurements of the detector with the laser (5 µs delay and 200 µs detector gate width), a digital pulse generator was used.

For the experiments, the tissues of 6 domestic pigs, 6 months of age on average, were used. The sagittal split heads were obtained from the local slaughterhouse on the day of slaughter. All animals were free from local or systemic diseases. The ex-vivo time before the experiments was kept within a time-frame of 6 h to minimize alterations due to protein degradation or exsiccation. For the nerve tissue, the infraorbital branch of the trigeminal nerve (N. V) with a caliber of approximately 655

Plasma Science and Technology, Vol.18, No.6, Jun. 2016

2.3

Measurement and damage assessment procedure

Before the LIBS measurements were performed, 3 laser pulses were sent to the focus spot on each tissue sample to remove any potential superficial contamination. Then, the LIBS spectra were collected from 6 different spots of each of the 6 gland and 6 nerve samples; 9 spectra were collected from each spot, yielding a total of 54 spectra from each tissue sample. After the measurements, one gland and one nerve tissue sample were examined for damage caused by the probing LIBS laser using an Optical Coherence Tomography system (Telesto II Optical Coherence Tomography, Thorlabs, Germany).

to calculate the intensity ratio in order to differentiate the tissues. Even though some elements show a marked difference in their mean intensity value, the higher standard deviation of the spectra makes it difficult to rely on visual inspection. Therefore, a systematic approach using PCA is taken to select the elements that contribute most to the variance in the data. Here, the representation of the data using the new variables (PCs) containing the majority of the variance reveals the contribution of each of the elements to the variance. Hence, the first 6 most contributing emissions were pursued for further analysis of differentiation using ratios by pairing these emissions in different combinations. The differentiation performance, sensitivity and specificity, of each of the selected pairs are evaluated using ROC.

2.4

3

Statistical analysis

For the statistical analysis, the LIBS spectra collected during the experiments were first normalized and mean centered. To achieve dimensionality reduction, a Principal Component Analysis (PCA) was performed by grouping the measurements as either training data or sample data. Here, each time, the data of 5 animals were used as training data to generate new variables (Principal Components) with reduced dimensionality. The data of the remaining animal (sample data) were then projected on them. This ensured that the principal components (PCs) used for the tissue classification were independent of the sample data. During the PCA, it was found that the majority of the data’s variance is contained in the first 19 PCs. All of the following analyses therefore were based on the data represented by those new variables. To achieve the tissue discrimination, a multiclass Linear Discriminant Analysis (LDA) was performed on the data resulting from the PCA, using six fold leaveone-out cross-validation. Each time, 5 animals were again used as training data and the remaining one animal as the sample data, resulting in a prediction of tissue class membership for all of the measurements. The performance of the classifier was then evaluated using the Receiver Operating Characteristic (ROC) by calculating the Area Under the ROC Curve (AUC). Additionally, the sensitivity and specificity values at the cut-off point were determined. All statistical analyses were performed using the Matlab (R2014a, MathWorks, Natick, USA) statistical toolbox.

2.5

Results and discussion

The mean spectra (Mean spectra and Standard Deviation) of the investigated gland and tissue samples of animal #1 are shown in Fig. 2 (54 spectra each). The experiments on the samples of the other animals yielded similar results. The prominent elements (Fig. 2) common to both tissue types are carbon (C), magnesium (Mg), iron (Fe), calcium (Ca), sodium (Na), hydrogen (H), nitrogen (N), potassium (K) and oxygen (O). A molecular emission from CN molecules is also observable in the spectra of both tissues. Upon looking at the spectra with the naked eye, already, obvious differences between the spectral intensity of certain elements can be seen, indicating that the concentrations of those elements differ in the two tissue classes.

Identification of elements and intensity ratio analysis

After processing the collected raw data, the elements responsible for the emissions at the different wavelengths were determined using the atomic emission database of the “National Institute of Standards and Technology” (NIST, Gaithersburg, USA) [35] . First, the mean spectra of the two tissue classes were inspected visually to find the elements that can be paired

Fig.2 Mean LIBS spectra of Gland (A) and Nerve tissue (B) of animal #1

656

F. MEHARI et al.: Investigation of LIBS for the Differentiation of Nerve and Gland Tissue Based on this, two approaches for the differentiation of the tissues can be taken: based on statistical analysis techniques or based on the intensity-ratio analysis between certain pairs of elemental emission lines, as described above. The intensity counts of carbon (C at 247.85 nm), magnesium (Mg at 279.55 nm), calcium (doublet at 393.52 nm and at 396.84 nm) and the carbon related CN emission (band-head near 388 nm) show distinct differences between the tissues and are possible candidates for the differentiation approach using the ratio-value analysis. A slight difference of the sodium (doublet at 588.99 nm and 589.59 nm) and oxygen (triplet near 777 nm) intensities is also observed.

3.1

discrimination, a multiclass LDA was performed using six-fold leave-one-out cross validation on the 6 animals. The resulting confusion matrix is shown in Table 2. A high classification performance can be achieved; there was no misclassification in all of the cases as the classifier was able to accurately differentiate all nerve tissue measurements from the respective gland tissue measurements.

Tissue discrimination by statistical analysis and ratio analysis

As mentioned above, a PCA was performed to reduce the number of variables to 19 main principal components (PCs) while representing the original spectra without much loss of information. Table 1 shows the portion of the variance contained in PC1, PC2 and PC1-PC19. Of all of the training datasets, an average of 35.39% of the total variance is contained in PC1. PC2 accounts for an average of 21.07% of the total variance. Overall, the reduction of the dimensionality of the dataset to 19 PCs covers an average of 83.28% of the variance in each of the training datasets. Summarizing the whole multidimensional dataset in only 19 newly derived variables is sufficient to differentiate the two tissue types with high accuracy.

Fig.3 (A) Scatter plot of data from 5 animals used as training dataset along PC1 and PC2. (B) Scatter plot of animal #6 sample data along PC1 and PC2 derived from the training dataset

Table 1. Percentage of variance contained in the first 19 PCs of the training datasets Training dataset

PC1

PC2

PC1-PC19

Animals Animals Animals Animals Animals Animals

34.52 37.55 36.66 34.28 34.21 35.14

19.73 20.51 19.63 22.67 22.19 21.7

82.29 83.43 83.56 83.62 83.67 83.11

1,2,3,4,5 1,2,3,4,6 1,2,3,5,6 1,2,4,5,6 1,3,4,5,6 2,3,4,5,6

Table 2. Confusion matrices of 6 animals obtained from LDA analysis using 19 PCs

As an example of the PCA, the scatter plot of the training data (representing animals 1-5) is shown in Fig. 3(A). The ability of the PCs to differentiate between the tissue classes can however best be evaluated when another dataset, the sample data (representing animal 6), not used for deriving the PCs, are projected on the new variables derived from the training data (Fig. 3(B)). The result shows a clear separation of the classes. In this manner, the datasets of all 6 animals were analyzed according to the training data derived from the remaining 5 animals. Then, to achieve tissue Table 3.

Animal 1 Gland Nerve

Gland 54 0

Nerve 0 54

Animal 2 Gland Nerve

Gland 54 0

Nerve 0 54

Animal 3 Gland Nerve

Gland 54 0

Nerve 0 54

Animal 4 Gland Nerve

Gland 54 0

Nerve 0 54

Animal 5 Gland Nerve

Gland 54 0

Nerve 0 54

Animal 6 Gland Nerve

Gland 54 0

Nerve 0 54

The performance of the classifier was further evaluated with ROC. The analysis again shows a high accuracy. The AUC value of the ROC graphs is 1 in all of the investigated animals. Equally, the sensitivity and specificity values at the cut-off point are 100% for each of the animals (Table 3).

AUC, sensitivity and specificity values of Nerve and Gland tissue classifications of 6 animals

Animal 1 Sensitivity (%) Specificity (%)

Animal 1 100 100

Animal 2 100 100

Animal 3 100 100

657

Animal 4 100 100

Animal 5 100 100

Animal 6 100 100

Plasma Science and Technology, Vol.18, No.6, Jun. 2016 For the ratio-value differentiation approach, as already mentioned above, a visual inspection of the spectra from the two tissues shows significant difference in the intensity of the emissions of carbon (at 247.85 nm), magnesium (at 279.55 nm), calcium (at 393.52 nm and at 396.84 nm) and carbon related CN emission (bandhead near 388 nm). Before these visually selected peaks are directly used for the differentiation, their significance in the separation of the dataset is validated by analyzing the PCs obtained during the PCA. The projection of the dataset on PC1 and PC2 (Fig. 3(A) and (B)) shows that the two classes are well separated along PC1. The figure also shows that the separation is further aided by PC2. Following this observation, the loadings of PC1 and PC2 were examined to identify how much the chosen emissions contribute to each PC (Fig. 4). Here, the positive and negative values on the ordinate indicate the direction that a given wavelength takes up relative to the mid-point (zero-value) of the given one dimensional vector (PC). The absolute value indicates the impact of each wavelength. Here, the same emissions first selected after visual inspection of the spectra (carbon, magnesium calcium and carbon related CN) are among the most-contributing emissions to PC1. Additionally, emissions related to sodium (doublet at 588.99 nm and 589.59 nm) and oxygen (triplet near 777 nm) are found to be among the most contributing emissions to PC2.

Considering only one emission line from the atoms with doublet and triplet emission wavelengths, the aforementioned 6 dominant emitting species can be paired in 15 different combinations to generate ratio values. The differentiation performance of each pair is then evaluated using ROC analysis. The sensitivity and specificity values of the pairs at the cut-off points are summarized in Fig. 5. The sensitivity values (Fig. 5(A)) show that the pairs C:Mg, Mg:CN, Mg:Ca, CN:Na and CN:O allow for a very good differentiation performance (more than 95% sensitivity on average). Especially C:Mg and Mg:CN show 100% correct differentiation in all of the investigated animals. The same is true for the specificity values (Fig. 5(B)). These results indicate that the simple ratio-value approach may be used as an alternative to the statistical analysis using the original variables (wavelengths) without compromising the accuracy of the results. It also appears that, as proven by the analysis of the PC loadings, the correct ratio-combinations may be chosen by simple visual inspection of the mean spectra of the investigated tissue pair.

Fig.5 (A) Sensitivity values of using intensity ratio of pairs of emissions for differentiation, (B) Specificity values of using intensity ratio of pairs of emissions for differentiation

It is worth mentioning that although single-emission lines appear multiple times in the selected emission pairs (e.g. Mg and CN), each of these lines has not been considered as an internal standard. Finding an internal standard proved difficult, as the measurement uncertainty was high even among the spectra of the same tissue type. This can mainly be attributed to the heterogeneous nature of the tissues and is, for example, manifested by the high variation in the intensity

Fig.4 (A) Mean plot of 6 PC1 loadings of original variables (wavelength) from 6 training datasets, (B) Mean plot of 6 PC2 loadings of original variables from 6 training datasets

658

F. MEHARI et al.: Investigation of LIBS for the Differentiation of Nerve and Gland Tissue of the hydrogen emissions. Hydrogen is present in soft tissues with similar concentration levels [36] and would serve as a reference element provided that the tissues are homogenous. To check its use as an internal standard, analysis was done on all hydrogen related ratios (not included in the pairs above) and lower sensitivity values are obtained (less than 95%).

3.2

performance of the technique. Here, reducing the laser energy, using tighter focusing and using shorter pulse durations might be reasonable approaches. However, care must be taken when using very small laser focus sizes, especially when getting close to the size of primary tissue-structures (e.g. connective tissue strands in gland tissue), as this may result in complete misrepresentation of the results. The current results have proven the potential of LIBS in providing feedback information for laser-based tissue specific surgery and may be used as a foundation for in-vivo animal studies. The in-vivo studies are in turn expected to provide an insight into the challenges posed during real surgical procedures. Experiments under such conditions are expected to address the effects of the sample-contamination mainly from blood, and the versatility and robustness of the technique in a clinical environment.

Technical and medical considerations

To assess the damage caused by the LIBS probing laser, the ablated area was investigated with the help of an OCT. Although upon visual inspection with the naked eye, no ablation crater could be found, the OCT images shown in Fig. 6 reveal that the damage caused to gland tissue by a typical 10-pulse measurement results in a defect of around 300 µm in outer-diameter and 300 µm in depth. The damaged area found in nerve tissue turned out to be a bit smaller (approximately 200 µm in outer-diameter). In this context, a previous study was able to show that a thermal damage of approximately 250 µm in diameter to a 1 mm motor nerve had no negative effects on nerve function [27] . In both tissues, a single pulse was not able to cause any damage recordable with the OCT. This suggests that the non-ideal but unavoidable damage caused by the probing LIBS laser may be tolerable. Additionally, although subject to further investigation, owing to the short pulse duration of the nanosecond pulsed Nd:YAG laser used in the present study, thermal effects around the ablation crater are not expected to exceed the size of the crater causing functional damaging effects to the tissues.

4

Conclusions

The present study successfully demonstrated LIBS based differentiation between porcine gland and nerve tissue. Both, a statistical approach using PCA and LDA, as well as the approach using element-specific intensity ratio comparison yielded good differentiation performances of 100% sensitivity and specificity. The results suggest that LIBS can reliably provide real time information on the investigated biological samples. The approach may be a candidate for the development of a real time feedback control mechanism for surgical laser systems that prevents the unwanted removal of certain tissues.

Acknowledgments The authors gratefully acknowledge the funding of the Erlangen Graduate School in Advanced Optical Technologies (SAOT) by the Deutsche Forschungsgemeinschaft (German Research Foundation - DFG) within the framework of the Initiative for Excellence.

References 1 2 3 4 5 6

Fig.6 Optical Coherence Tomography (OCT) images of the ablation craters from 10 LIBS laser pulses from Gland tissue and Nerve tissue

7 8 9 10

One of the future goals in LIBS-based tissue differentiation should be to further optimize the experimental parameters in order to reduce the laser energy deposited on the tissues, without compromising the differentiation 659

Iro H, Zenk J. 2014, Laryngo Rhino Otol., 93: s103 Foresta E, Torroni A, Nardo F, et al. 2014, Oral Surg. Oral Med-Oral Pathol. Oral Radiol., 117: 663 Lagha A, Chraiet N, Ayadi M, et al. 2012, Head Neck Onco., 4: 19 Laurie S A, Licitra L. 2006, J. Clin. Oncol., 24: 2673 Gosain A. 1995, Clin. Plast. Surg., 22: 241 Davis R, Anson B, Budinger J, et al. 1956, Surg. Gynecol. Obstet., 102: 385 Tweedie D, Jacob A. 2009, Clin. Otolaryngol., 34: 303 Witt R. 2002, Laryngoscope, 112: 2141 O’Brien C. 2003, Head Neck, 25: 946 Iro H, Zenk J, Koch M, et al. 2013, Head Neck, 35: 788

Plasma Science and Technology, Vol.18, No.6, Jun. 2016 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

Nouraei S, Hope K, Kelly C, et al. 2005, Plast. Reconstr. Surg., 116: 1206 Ho A, Scott A, Klassen A, et al. 2012, Plast. Reconstr. Surg., 130: 91 Carruth J. 1984, J. Med. Eng. Technol., 8: 161 Ben-Bassat M, Kaplan I, Shindel Y, et al. 1978, Br. J. Plast. Surg., 31: 155 Strong M, Vaughan C, Healy G, et al. 1979, Laryngoscope, 89: 897 Carruth J. 1982, J. Laryngol. Otol., 96: 529 Gaspar L. 1994, J. Clin. Laser Med. Surg., 12: 281 Schoelch M, Sekandari N, Regezi J, et al. 1999, Laryngoscope, 109: 949 Astor F, Hanft K. 2003, J. Clin. Laser Med. Surg., 21: 297 Basu M, Frame J, Rhys Evans P. 1988, J. Laryngol. Otol., 102: 322 Luomanen M. 1987, Eur. J. Oral Sci., 95: 65 Fisher S, Frame J. 1984, Br. J. Oral Maxillofac. Surg., 22: 414 Bryant G, Davidson J, Ossoff R, et al. 1998, Laryngoscope, 108: 13 Lopez-Jornet P, Camacho-Alonso F. 2013, Med. Oral Patol. Oral Cir. Bucal., 18: e38 Baxter G, Walsh D, Allen J, et al. 1994, Exp. Physiol., 79: 227 Menovsky T, Van Den Bergh Weerman M, Beek J. 1996, Microsurgery, 17: 562

27 28 29 30 31

32 33 34

35

36

Menovsky T, Van Den Bergh Weerman M, Beek J. 2000, Microsurgery, 20: 150 Orchardson R, Peacock J, Whitters C. 1997, Lasers Surg. Med., 21: 142 Kanawade R, Mehari F, Knipfer C, et al. 2013, Spectrochim. Acta Part B: At. Spectrosc., 87: 175 Mehari F, Rohde M, Knipfer C, et al. 2014, Biomed. Opt. Express, 5: 4013 Cremers D. 2006, Handbook of Laser-Induced Breakdown Spectroscopy. John and Wiley, Chichester, England Cremers D, Chinni R. 2009, Appl. Spectrosc. Rev., 44: 457 Myakalwar A, Sreedhar S, Barman I, et al. 2011, J. Talanta, 87: 53 Samek O, Liska M, Kaiser J, et al. 1999, Laser ablation for mineral analysis in the human body: integration of LIFS with LIBS, SPIE, Stockholm, Sweden, p.263 Kramida A, Ralchenko Y, Reader J. 2012, NIST Atomic Spectra Database (ver. 5.0), National Institute of Standards and Technology, Gaithersburg, MD (Available at: http://physics. nist. gov/asd) Woodard H, White D. 1986, Br. J. Radiol., 59: 1209

(Manuscript received 1 September 2015) (Manuscript accepted 21 October 2015) E-mail address of F. MEHARI: [email protected]

660