H+, O2+, O3+ and high resolution PIXE spectra of Yb2O3

Nuclear Instruments and Methods in Physics Research B 410 (2017) 193–199

Contents lists available at ScienceDirect

Nuclear Instruments and Methods in Physics Research B journal homepage: www.elsevier.com/locate/nimb

H+, O2+, O3+ and high resolution PIXE spectra of Yb2O3 P.C. Chaves a,⇑, M.A. Reis a,b a b

Centro de Ciências e Tecnologias Nucleares (C2TN), Instituto Superior Técnico, Universidade de Lisboa, Campus Tecnológico e Nuclear, EN10 km 139.7, 2695-066 Bobadela, Portugal IEQUALTECS, Lda, R. Dr. Francisco Sá Carneiro, 36 2500-065 S. Gregório CLD, Portugal

a r t i c l e

i n f o

Article history: Received 17 May 2017 Received in revised form 19 July 2017 Accepted 19 August 2017

Keyword: XMS HRHE-PIXE DT2 Multiple-ionizations Rare Earth Elements

a b s t r a c t The number of X-ray spectrometry systems having energy resolution of the order of 10 eV, or less, has increasing recently, included already energy dispersive systems (EDS). Access to previous unseen spectra details and enhanced information including speciation, becomes more common and available. Analysis of high resolution EDS PIXE spectra is, nevertheless a complex task due to the need to carefully account for contributions from minor and satellite transitions. In this work, a pure Yb2 O3 sample was irradiated at the HRHE-PIXE setup of C2TN, and simultaneous CdTe and X-ray Microcalorimeter Spectrometer (XMS) spectra were collected. The L-shell spectrum of Yb emitted during irradiations using Hþ , O2þ and O3þ ions in the energy range from 1.0 to 6.5 MeV was studied. Measured L X-ray spectra were analysed taking into account the effects of the multiple ionization in the L and M shells. All spectra were analysed using the DT2 code, which allows to include in the fitting model diagram lines as well as multi-ionization satellites and any other contributions. In this communication we present the results and discuss details and problems related to the transition energies, intensity, line width data, and multiple ionization satellites. Ó 2017 Elsevier B.V. All rights reserved.

1. Introduction Particle induced X-ray emission (PIXE) is a powerful technique for quantitative analysis because it is non-destructive, multielemental (from Na to U), highly sensitive and requires no special sample preparation. Usually proton beams with an energy around 2 MeV are used in PIXE offering high sensitivity. Since the sensitivity of PIXE strongly depends the ionization cross-sections, which are proportional to the square of the projectile atomic number, it can be expected that the use of heavy ion beams improves the sensitivity of the analysis considerably. However, the X-rays emitted from atoms ionized by heavy ions exhibit, apart from the X-ray diagram lines, a significant satellite structure corresponding to different multi-vacancy configurations. Heavy ions PIXE is still far from being an established analytical method and various issues such as the determination of fluorescence yield, line shifts and peak broadening due to multiple ionization, still need to be properly established. In Fig. 1 the overlap of three spectra of thick Yb2O3 targets irradiated using different ion beams is shown. In this case, spectra obtained in the lowest standard stable energy conditions of the used accelerator are presented. The complexity of the analysis becomes clear from the differences put in evidence. In fact, even if the differences in Lb to La groups could be explained by matrix ⇑ Corresponding author. E-mail address: [email protected] (P.C. Chaves). http://dx.doi.org/10.1016/j.nimb.2017.08.026 0168-583X/Ó 2017 Elsevier B.V. All rights reserved.

effects, the different shapes of the Lb group cannot. Furthermore, since the physics of oxygen beam ionizations is more complex then that of proton beams the result is not unexpected. The differences in shape may be explain by differences in L2 and L3 ionization cross sections since the Lb group is composed by transition to both sub-shells. Nevertheless, high resolution studies in this area can enhance the gathering of information about the structure of multi-vacancy configurations, helping to extract information present in data collected by other detector systems. In this work the characteristics of the C2TN High Resolution High X-ray Energy (HRHE)-PIXE [1,2] and the DT2 [3] code versatility will be shown to be a major support to the study of these spectra. 2. Materials and methods 2.1. Experimental chamber and ion beams Experiments were carried out using the High Resolution High Xray Energy (HRHE)-PIXE setup [1] endstation of the CTN 3.0 MV Tandetron. X-rays were collected using an Amptek Peltier Cooled 3  3  1 mm3 CdTe detector placed at about 25 mm from target and at 145 relative to the beam direction. Detector window is a 250 lm Beryllium window. EDS high resolution L-shell spectra were collected simultaneous to the CdTe ones, using a Vericold Tech. GmbH Polaris XMS, referred here as the C2TN XMS, set at 90 to beam direction. More details about the HRHE-PIXE system

194

P.C. Chaves, M.A. Reis / Nuclear Instruments and Methods in Physics Research B 410 (2017) 193–199

dent on the rl ðE0 Þ. The spectra line Gaussian standard deviation is therefore, one of the most important parameters to take into account for a detailed fitting process. Fine adjusting its energy dependence for the present case, was made considering the Yb2 O3 spectrum collected by the CdTe detector when the sample was irradiated using 3.8 MeV Hþ ions. As shown in Fig. 2 (top), in this conditions K and L lines are present in spectrum so that an energy range from 7 to 61 keV can be used. The spectrum was fitted in two regions considering separately K and L X-ray lines. Based on the values for rl ðE0 Þ of main lines, La1 , Lb1 , Lc1 , Ka1 and Kb1 it was possible to establish the energy dependence of the spectra line Gaussian standard deviation, modeled as:

rl ðE0 Þ ¼ rcoef  ðasig þ bsig  E0 þ csig  E20 Þ

Fig. 1. Overlap of spectra obtained for Yb2 O3 when irradiated the sample with 3.0 MeV O2+ and O3+ ion beams with the Hþ spectrum collected when sample was irradiated with 1.0 MeV Hþ ion beams. Spectra were normalized to Lc peak. It can be seen an effect in the low energy region of O3+ spectrum not observed in the O2+ spectrum.

ð3Þ

where the asig, bsig and csig parameters must be carefully determined. The parameter rcoef is adjustable during each spectrum fitting procedure. In Fig. 2 (bottom) the experimental data and fit for the ‘‘Sigma”, rl ðE0 Þ energy dependence are presented. 2.3. XMS analysis of Yb transition energies, rates and widths

can be found in references [1,2]. A thick pellet of pure Yb2 O3 (99.998%) target was prepared from powder by pressing to 5.6 tons/cm2 and studied using ion beams of Hþ , O2þ and O3þ having energies from 1.0 to 6.5 MeV. Experimental condition details are presented in Table 1. Ion beam spot in the target was set to a circus of 3 mm in diameter by using a dual collimator providing an elliptical shaped beam cross-section [1]. An electron gun was used to avoid target charging up. In total, 8 spectra were collected. In the case of the irradiation using Hþ ions at 3.8 MeV, a funny filter of a 1 mm Al thick foil having a hole of 1 mm in diameter in the center, was used in front of the detector to level the detector efficiency over its wide dynamical range (5–120 keV). Beam currents were adjusted to eliminate the possibility of exist charge effects during the irradiation. 2.2. Spectra lines Gaussian standard deviation energy dependence The semiconductor detector response function is modeled as [4]:

FðEÞ ¼ GðEÞ þ CaðEÞ þ Ge ðEÞ

ð1Þ

where

h i2 EE  pffiffiffiffi 0 1 2prl ðE0 Þ GðEÞ ¼ pffiffiffiffiffiffiffi  rl ðE0 Þe 2p

ð2Þ

rl ðE0 Þ being the spectra line Gaussian standard deviation. The other two terms are relative to tail and escape peaks also depen-

Once set the rl ðE0 Þ parameter variation in energy, the second major issue is knowing the exact energy of all transitions present in the spectrum. Usually, the ratio of intensities is also required, but the DT2 code overcomes this difficulty when used in free line mode [3]. As mention above, in heavy ion PIXE, multiple ionization components are significant. Knowing their location is thus a major issue. In Fig. 3, the La energy region of the XMS high resolution spectrum of the Yb2 O3 target irradiated using a 1.0 MeV Hþ beam is shown. Fit and partial contributions to this, namely La1 , La2 and La1 multi ionizations are also presented. Transition energies for diagram lines were obtained from Bearden [5] and multiple ionizations transitions were calculated from these based on the work of Uchai et al. [6]. As can be seen, in this L X-ray energy range the background is close to zero and X-ray lines are well separated allowing the identification of La multi ionizations and making it possible to obtain a good fit. In Fig. 4 the Lb energy region is shown. In this region it is possible to identified the Lb4 , Lb1 , Lb6 , Lb3 and Lb2 X-ray lines. Single and double spectator vacancies were considered and fitted for Lb1 . In this fit the Lb6 line was excluded due to the energy proximity (8.6 eV) to the double spectator vacancy satellite of Lb1 , L2M4m2, combined with the fact that Lb6 has low intensity. High resolution spectra lines are modeled by DT2 using a true Voigt function. Based on the fit of the XMS spectrum presented in Figs. 3 and 4 spectrum, it was verified that the natural width values reported by Perkins and co-workers [7] for Lb1 and Lb2 are not

Table 1 Description of all experimental conditions used in the present work. XMS and CdTe specra collected at 1.0 MeV using proton beams were collected using the detectors simultaneously. Yb2 O3 spectrum collected at 3.8 MeV was used to study the sigma energy dependence. Detector

Particle

Beam energy (MeV)

Diffuser

Filter

Collimator (mm)

Charge (lC)

XMS

Hþ

1.0

Yes

No

3

90.6

CdTe

Hþ Hþ O2+ O2+ O3+ O3+ O3+

1.0 3.8 2.0 3.0 3.0 4.5 6.5

Yes Yes No No No No No

No BN + Al funny No No No No No

3 3 3 3 3 3 3

90.6 60 771 203 42 240 32

P.C. Chaves, M.A. Reis / Nuclear Instruments and Methods in Physics Research B 410 (2017) 193–199

195

Fig. 2. (a) Spectrum of Yb2 O3 pure thick sample obtained with the CdTe detector when irradiated at 3.8 MeV Hþ ion beam. In the right side of this figure a zoom (30) of the higher energy part (12–65 keV) are shown. (b) rl ðE0 Þ experimental data for La1 up to Kb of Yb (7-61 keV) lines. Fit was made and the values obtained were included in the DT2 code database [3] and used in the Yb2 O3 fits.

Fig. 3. (a) La spectrum region of Yb2 O3 pure thick sample obtained with the XMS detector when irradiated at 1.0 MeV Hþ ion beam. Spectrum was fitted using the DT2 code [3]. Single and double spectator vacancies can be observed for La1 Intensity ratios L3M5m1/L3M5 and L3M5m2/L3M5m1 were extract and used in CdTe Hþ spectrum fit.

compatible with the experimental data obtained in this work. Perkins and co-workers [7] reported natural widths of 6.56 and 9.98 eV, respectively, while our results require these natural widths to be of 20 eV for each of these lines, in order to obtain a proper fit. These new values were considered in the fits made after-

Fig. 4. Lb spectrum region and fit of Yb2 O3 pure thick sample obtained with the XMS detector when irradiated at 1.0 MeV Hþ ion beam. Single and double spectator vacancies were considered for Lb1 , however only single spectator vacancy is observed. From this spectrum/fit we extract the ratio L2M4m1/L2M4 and use this value in the following CdTe Hþ spectrum fit. Apart from this we verify that values reported by Perkins [7] for Lb1 and Lb2 X-ray lines are not in agreement in the experimental data and in this fit we already corrected them, values in bold in Table 3.

words. In Fig. 4 the fit presented already considered natural widths of the 20 eV.

196

P.C. Chaves, M.A. Reis / Nuclear Instruments and Methods in Physics Research B 410 (2017) 193–199 Table 2 Intensity ratios of multiple ionization to the parent lines obtained considering the XMS high resolution spectrum fit. Intensity ratio

Value (%)

L3M5m1/L3M5 L3M5m2/L3M5m1 L2M4m1/L2M4

10.8  0.3 22.5  0.7 14.2  0.1

In Table 2 the intensity ratios of multiple ionization to the parent lines obtained in this fit, are presented. 2.4. Yb transition energies and model The 1.0 MeV Hþ CdTe spectrum was collected for significant statistics (500 K counts in La1 ). A detailed list of energies is necessary for a proper fit. The list of X-ray energies used to fit K and L spectra of Yb2 O3 is presented in Table 3. Yb K and L X-rays lines are identified using both the Siegbhan and IUPAC notation. The vacancy transitions, the energies, energy difference between diagram lines, theoretical intensities and Xray lines widths are also present in Table 3. X-ray lines are ordered by energy (in keV). The values used were obtained from the Handbook of X-ray Data [8], mostly bearing values traceable to Bearden [5]. In what concerns the intensity of diagram lines, values reported by Scofield were used [9], and for the natural line widths (in eV) of the diagram lines the values reported by Perkins [7] were considered. In the case of multi-ionizations the energies were obtained from the work of Uchai and co-workers [6] for most of the cases. DT2 fit was carried out mostly in free mode [3], meaning that the lines intensities were not restricted to the intensity of any other lines. This was nevertheless not the case for all lines, in particular for the multiple ionization components. While in the XMS spectrum only multiple ionization components of the main lines were considered (search for ”U” in Table 3) being kept free, this was not the case for the CdTe spectra. Since the detection angles for both detectors in the HRHE-PIXE chamber are similar we considered that the three intensity ratios (Table 2) obtained considering the XMS spectrum, could be used to fit the spectrum collected with the CdTe detector, when the sample was irradiated using 1.0 MeV Hþ ion beams. These intensity ratios, check Table 2, were thus fixed during the CdTe spectrum fit. In what concerns other multiple ionizations transitions, single spectator vacancy satellites, m1, were set ad hoc by the authors to a maximum of 25% relative to the intensity of the main line, and double spectator vacancy satellites, m2, were set to a maximum intensity of 25% relative to the multiple ionizations m1. Observing preliminary fitting results, it was found that cases exist where transitions ought to be excluded from the model due to their extreme proximity to more intense ones. An energy difference of 20 eV (slightly more than 1 channel) between X-ray lines was set as minimum difference to accepted a less intense transition in the model. In Table 3 we present the final list of transitions energies used, excluded lines being identified as ‘‘exc.” in the last column. In the case of multiple ionization transitions observable in the CdTe detector but not in the XMS spectrum (due to lack of statistics) intensity ratios to the main line had to be determined without support of the high resolution spectrum data. In Table 3 the final energy line value for the lines reported by Uchai [6] and co-workers are presented in bold. Based on the values reported in their work (L3M1-Ll , L3M4-La2 , L3M5-La1 , L2M1-Lg ,

L2M4-Lb1 , L2N4-Lc1 , L1N2-Lc2 , L1N3-Lc3 , L1O2-Lc40 , L1O3-Lc4 ) values were estimated for multi ionization associated to other L X-ray transitions (all except the lines indicated as excluded in the last column, see Table 3). In the case of transitions from shells that are not reported by [6] the following decisions were adopted: for transitions from M2, energy shift values of transitions from M1 (L3M1) were selected, while for transitions from M3, values from M4 transitions (L2M4) were used. In the case of transitions from N1, the values for N2 shell (L1N2) were adopted and finally for transitions from N5 the values reported for N4 (L2N4) were chosen. In other cases the values reported for transitions to the same sub-shell were assumed. Multiple ionization transition line widths were considered to be identical to the main line for the case of m1, the double of this in the case of m2 and four times, in the case of L3M5m34. A non-identified line at 9.3 keV was included in the final fit process to compensate for a major uncompensated signal in residuals. Spectra deconvolutions were carried out using the revised version 0v9_27 of the DT2 code [3]. To avoid errors by the user during the data analysis process, a single input model was used for each ion beam experiment, thus only to 3 different input files were used. 3. Results and discussion In Fig. 5, a spectrum of Yb2 O3 collected using the CdTe detector during irradiation using a 1.0 MeV Hþ beam is shown. Residuals from the fit are presented underneath. A god fit was achieved, as is clear from the residuals. Some small nonaccounted contributions can still be inferred in energy regions were diagram transitions are absent. A detail zoom of the Lb energy region is presented in Fig. 6, in order to shown the complexity of the fit. The four main lines in this energy region, Lb4 , Lb1 , Lb3 and Lb2 plus the two first multi ionizations of each line were fitted and can be seen. The quality of the fit is clear from the residuals. In Fig. 7, a spectrum of the Yb2 O3 irradiated using a 4.5 MeV O3+ beam is shown. Residuals are once again presented underneath. In this case, all the lines mentioned in Table 3 as being use, were fitted in free mode. Once fitted all spectra from the Hþ , O2+ and O3+ experiments, peak area values for diagram and multi ionizations corresponding to L3, L2 and L1 transitions were added to determined the total sub-shell yield for each case. The ratios of areas obtained for oxygen beam irradiations to the corresponding areas from the spectrum collected during Hþ ion beam irradiation were calculated taking into account the collecting charge for each single spectrum, corrected to the ion charge. In Fig. 8 the results for these ratios are presented as function of the ion beam energy. It can be seen that, at 3.0 MeV the results for the O2+ irradiation and the results for the O3+ ion beam are not identical, and uncertainties do not compensate the difference, opposed to what could be expectable. In the case of L3 and L2 sub-shells the values obtained for the O2+ case being nearly twice those for O3+. In other to try to understand the reasons for these results an analysis of Fig. 1 shows that the spectra for 3.0 MeV O2+ and for O3+ are very similar except in the low energy tailing of the two major groups La and Lb . Not even taking into account the low collecting charge of the O3+ spectrum is it possible to overcame the differences observed, since the noise in the tail counts is much smaller than the differences in the ratio from La peak to the tail. This implies that the O3+ beam induces processes within the target, witch are different from those induced by O2+. These ones being more similar to Hþ in what concerns to low energy tailing signals. Furthermore, one can speculate that transitions after O3+

197

P.C. Chaves, M.A. Reis / Nuclear Instruments and Methods in Physics Research B 410 (2017) 193–199

Table 3 Vacancy transitions, transition energies [5], energy difference between subsequent lines, intensity and widths of the X-ray lines. Yb X-ray lines are described using the Siegbahn notation and IUPAC notation. Escape peaks were considered in the detector response function. In the table, numbers in brackets mean powers of 10, eg: (2) means x10ð2Þ and the following legend applies: (a) energy values assumed by the authors taking into consideration values reported by Uchai [6] for similar transitions; (b) initial values for the intensity of multi-ionizations were set ad hoc by the authors to a maximum of 25% relative to the intensity of the main line and double spectator vacancy satellites, m2, were set to a maximum intensity of 25% relative to the multiple ionization m1. (c) multi ionization transitions energy widths were considered to be identical to that of the main line for the case of m1, the double of this in the case of m2 and four times in the case of L3M5m34; (d) energy width assumed by the authors taking into consideration the XMS fit (values in parentheses corresponds to the previous values reported by [7]); (e) intensity values assumed by the authors when no values were found in the literature. Finally in the last column we identified the lines considered in the XMS spectrum fit and also the lines that was initially considered but excluded form Oxygen spectra fits due to problems related with overlaps in energy or very low intensity. These lines are identified as ‘‘exc.” Sieg.

IUPAC

Vacancy transitions

Ener. (keV)

Ka2

KL2

2p1=2 ! 1s1=2

51.3546

Ka1

KL3

2p3=2 ! 1s1=2

52.3895

Kb3

KM2

3p1=2 ! 1s1=2

Kb1

KM3

3p3=2 ! 1s1=2

Kb2

KN23

4p1=2;3=2 ! 1s1=2

Ll

L3M1

3s1=2 ! 2p3=2

6.5484

Lt

L3M1m1 L3M1m2 L3M2

3s 3s 3p1=2 ! 2p3=2

Ls

L3M2m1 L3M2m2 L3M3

DE (eV)

Ener. Ref.

Intensity (eV/ h)

Width (eV) [7]

[5]

8.67(+0)

35.54

–

1034.9

[5]

1.528(+1)

35.48

–

59.1520

6762.5

[5]

1.666(+0)

41.59

–

59.3671

215.1

[5]

3.22(+0)

41.77

–

60.9850

1617.9

[5]

3.55(1)

45.00

–

[12]

3.42(2)

22.83

U

6.5754 6.6034 6.771

27 28 167.6

[6] [6] [12]

(b) (b) 3.27(4)

22.83 (c) 45.66 (c) 14.46

U

3s 3s 3p3=2 ! 2p3=2

6.798 (a) 6.825 (a) 6.9942

27 27 169.2

[6] [6] [12]

(b) (b) 2.98(4)

14.46 (c) 28.92 (c) 14.64

U

La2

L3M3m1 L3M3m2 L3M4

3s 3s 3d3=2 ! 2p3=2

7.0172 (a) 7.0402 (a) 7.3691

23 23 328.9

[6] [6] [12]

(b) (b) 7.84 (2)

14.64 (c) 29.28 (c) 6.48

U

La1

L3M4m1 L3M5

3p 3d5=2 ! 2p3=2

7.3881 7.4163

19 28.2

[6] [12]

(b) 6.93(1)

6.48 (c) 5.48

U U

Lg

L3M5m1 L3M5m2 L3M5m34 L2M1

3p 3p 3p 3s1=2 ! 2p1=2

7.4353 7.4563 7.4888 7.5828

19 21 32.5 94

[6] [6] [6] [12]

(b) (b) (b) 2.325(2)

5.48 (c) 10.96 (c) 21.92 (c) 22.91

U U – U

L2M1m1 L2M1m2 L2M2

3s 3s 3p1=2 ! 9m1=2

7.6138 7.6458 7.8054

31 32 159.6

[6] [6] [12]

(b) (b) 4.36(7)

22.91 (c) 45.82 (c) 14.54

U

Lb17

L2M2m1 L2M2m2 L2M3

3s 3s 3p3=2 ! 2p1=2

7.8324 (a) 7.8594 (a) 8.0286

27 27 169.2

[6] [6] [12]

(b) (b) 6.76(4)

14.54 (c) 29.08 (c) 14.71

U

3s 3s 3s1=2 ! 2s1=2 3p1=2 ! 2s1=2

8.0516 (a) 8.0746 (a) 8.091 8.3136

23 23 16.4

Lb4

L2M3m1 L2M3m2 L1M1 L1M2

[6] [6] [12] [12]

(b) (b) 1.58(6) 1.967(1)

14.71 (c) 29.42 (c) 26.74 18.36

Uexc. U

Lb1

L1M2m1 L1M2m2 L2M4

3s 3s 3d3=2 ! 2p1=2

8.3406 (a) 8.3676 (a) 8.4036

27 27 36

– – [12]

(b) (b) 8.66(1)

18.36 (c) 36.72 (c) 20(6.56)(d)

U U U

Lb6

L2M4m1 L2M4m2 L3N1

3p 3p 4s1=2 ! 2p3=2

8.4266 8.4496 8.4582

23 23 8.6

[6] [6] [12]

(b) (b) 8.05(3)

20 (d,c) 20 (d,c) 9.788

U U Uexc.

Lb3

L2M4m34 L1M3

3p 3p3=2 ! 2s1=2

8.4836 8.5368

25.4 53.2

[6] [12]

(b) 2.607(1)

40 (d,c) 18.55

U U

Lb15

L1M3m1 L1M3m2 L3N4

3s 3s 4d3=2 ! 2p3=2

8.5598 (a) 8.5828 (a) 8.74373

23 23 160.93

– – [12]

(b) (b) (b)

18.55 (c) 37.10 (c) 11.39

– – Uexc.

Lb2

L3N5

4d5=2 ! 2p3=2

8.75569

11.96

[12]

1.182(1)

20(9.98)(d)

U

4p 4p 3d3=2 ! 2s1=2 3p 3p 3d5=2 ! 2s1=2 3p 3p

60 60 36.01 23 23 1.2 19 21 301.1 192.7

[12]

(b) (b) 5.07(3) (b) (b) 7.59(3) (b) (b) (b) 5.63(3)

20 (d,c) 40 (d,c) 10.39 10.39 (c) 20.78 (c) 9.39 9.39 (c) 18.78 (c) 10 (x) 17.96

U – U

4s1=2 ! 2p1=2

8.81569 (a) 8.87569 (a) 8.9117 8.9347 (a) 8.9577 (a) 8.9589 8.9779 (a) 8.9989 (a) 9.3 9.4927

[6] [6] [12] [6] [6] [12] [6] [6]

Lc5

L3N5m1 L3N5m2 L1M4 L1M4m1 L1M4m2 L1M5 L1M5m1 L1M5m2 other L2N1

Lc1

L2N1m1 L2N1m2 L2N4

3d 3d 4d3=2 ! 2p1=2

9.5447 (a) 9.5997 (a) 9.7782

52 55 178.5

[6] [6] [12]

(b) (b) 1.543(1)

17.96 (c) 35.92 (c) 11.46

L2N5

4d5=2 ! 2p1=2

9.79013

11.93

[12]

1.00(5) (e)

10.05

Uexc.

Lc2

L2N4m1 L2N4m2 L1N2

4p 4p 4p1=2 ! 2s1=2

9.8382 9.8952 10.0926

48.07 57 197.4

[6] [6] [12]

(b) (b) 4.72(2)

11.46 (c) 22.92 (c) 12.00

U U U

Lc3

L1N3

4p3=2 ! 2s1=2

10.142

49.4

[12]

6.56(2)

12.00

U

L1N2m1

4s

10.1446

2.6

[6]

(b)

12.00 (c)

exc.

Lb10

Lb9

XMS

Uexc.

U

U

(continued on next page)

198

P.C. Chaves, M.A. Reis / Nuclear Instruments and Methods in Physics Research B 410 (2017) 193–199

Table 3 (continued) Sieg.

IUPAC

Vacancy transitions

Ener. (keV)

DE (eV)

Ener. Ref.

Intensity (eV/ h)

Width (eV) [7]

Lc11

L1N3m1 L1N2m2 L1N3m2 L1N45 L1N45m1 L1N45m2 L1O23

4s 4s 4s 4d3=2;5=2 ! 2s1=2 4p 4p 5p1=2;3=2 ! 2s1=2

10.196 10.1996 10.252 10.2923 10.3523 (a) 10.4123 (a) 10.4603

51.4 3.6 52.4 40.3 60 60 480

[6] [6] [6] [12] [6] [6] [5]

(b) (b) (b) 1.00(5) (e) (b) (b) 6.16(3)

12.00 (c) 24.00 (c) 24.00 (c) 10.0 (c) 10.0 (c) 20.0 (c) 8.0

L1O23m1 L1O23m2

5s 5s

10.5253 10.5933

65 68

[6] [6]

(b) (b)

8.0 (c) 16.0 (c)

Lc4

Fig. 5. Spectrum and fit of Yb2 O3 pure thick sample obtained with the CdTe detector when irradiated at 1.0 MeV Hþ ion beam. The ratios values of L3M5m1/ L3M5, L3M5m2/L3M5m1 and L2M4m1/L2M4 obtained from XMS spectrum were imposed in this spectrum fit.

collisions have an enhancement in Auger decay processes following the ionization, which lead to the background shape enhancement in the O3+ spectrum. Taking the Z 2 =E4 rule one expects that 3 MeV oxygen beam has a yield that is 64  (3/16)4 = 0.08 of the 1.0 MeV Hþ spectrum, compatible with the results for the L2 sub-shell for O3+ case and L3 for the O2+ case but not for any of the other situations at 3.0 MeV oxygen beam experiments.

XMS exc. U

U

Fig. 7. Spectrum of Yb2 O3 pure thick sample obtained with the CdTe detector when irradiated at 4.5 MeV O3+ ion beam. Spectrum was fitted using the DT2 code [3]. All lines in Table 3 were fitted in free mode. The details can also be observed in the lower figure where fitting results is also present.

A change in the sub-shell fluorescence yields [10] may justify these differences. This is compatible with a higher Auger yield from the L3 sub-shell after O3+ collisions, which increases the continuous background around the La group, and a lower Auger yield from L2 sub-shell after O2+ collisions, which will reduce this same signal, as observed in Fig. 1. A similar explanation may be tentatively used to explain the much higher yield of the single spectator vacancy satellite compared to the diagram transition in the case of

Fig. 6. Lb spectrum region focusing in the energy region containing the Lb4 , Lb1 , Lb3 and Lb2 lines corresponding multi ionizations (see Table 3).

P.C. Chaves, M.A. Reis / Nuclear Instruments and Methods in Physics Research B 410 (2017) 193–199

199

Ma in Ta thin films irradiated with oxygen beams as recently reported elsewhere [11]. 4. Conclusions In this work we have shown that Yb L spectra induced by Hþ , O2+ and O3+ ion beams are all different, being that these differences are not explainable by matrix effects. High resolution work provided new values of 20 eV for Lb1 and Lb2 natural line widths, and main multiple-ionization satellite intensities, relative to their parent line. Lack of literature data on multiple ionizations energies, intensities and line widths is a fact, which must be overcome. Although a systematic databases on these might not be feasible, it is important to find a proper solution for the problem. In the present study, energy and line width values for several multiple ionizations transitions were adopted in an ad hoc way. These values should be confirmed after additional work. Finally, results for 3.0 MeV irradiations using O2+ beams and O3+ beam are not identical as expected. The detailed analysis of the spectra take us to propose different fluorescence yield in each case, still whether this is indeed the reason for the observations, must be confirmed by additional work. Acknowledgements This work was made possible by the partial financial support of the Portuguese Foundation for Science and Technology, FCT, fellowship SFRH/BPD/76733/2011 and through the UID/ Multi/04349/2013 project and to the IAEA support through Research contract No. 18357 in the frame of the Coordinated Research Project F11019 on Development of Molecular Concentration Mapping Techniques using MeV Focused Ion Beams. References

Fig. 8. Ratios of areas (peak area values for diagram and multi ionizations corresponding to L3, L2 and L1 transitions were added to determined the total sub-shell yield for each case) obtained for Oxygen beam irradiations to the corresponding areas from the spectrum collected using Hþ as function of the ion beam energy. It is taking into account the collecting charge for each single spectrum, corrected to the ion charge. As can be seen, at 3.0 MeV the results for O2+ case are nearly twice for O3+ case.

[1] P.C. Chaves, A. Taborda, J.P. Marques, M.A. Reis, N to K Uranium PIXE spectra obtained at the High Resolution High Energy PIXE setup, Nucl. Instr. Meth. Phys. Res. B 318 (2014) 60–64. [2] M.A. Reis, P.C. Chaves, A. Taborda, Review and perspectives on energy dispersive High Resolution PIXE and RYIED, Appl. Spectrosc. Rev. (2016). [3] M.A. Reis, P.C. Chaves, A. Taborda, J.P. Marques, N.P. Barradas, Fixed and free line ratio DT2 PIXE fitting and simulation package, Nucl. Instr. Meth. Phys. Res. B 318 (65–69) (2014). [4] M.A. Reis, P.C. Chaves, L.C. Alves, N.P. Barradas, DT2, a PIXE spectra simulation and fitting package, X-Ray Spectrom. 37 (2008) 100–102. [5] J.A. Bearden, X-ray wavelengths, Rev. Modern Phys. 39 (78) (1967). [6] W. Uchai, C.W. Nestor Jr., R. Raman, C.R. Vane, Energy shifts of L X-rays from 706z690 elements due to multiple M vacancies, Atom. Data Nucl. Data Tables 34 (1986) 201–214. [7] S.T. Perkins, D.E. Cullen, M.H. Chen, J.H.Hubbell, J. Rathkopf, J.H. Scofield, Tables and Graphs of Atomic Subshell Relaxation Data Derived from the LLNL Evaluated Atomic Data Library Z=1-100, vol. 30, 1991. [8] Günter Zschornack, Handbook of X-Ray Data, Springer, 2007. [9] J.H. Scofield, Relativistic Hartree-Slater values for K and L X-ray emission rates, Atom. Data Nucl. Data Tables 14 (1974) 121. [10] M.A. Reis, A. Jesus, Semiempirical approximation to cross sections for L X-ray production by proton impact, Atom. Data Nucl. Data Tables 63 (1996) 1–55. [11] M.A. Reis, P.C. Chaves, CdTe and EDS HR-PIXE Ta L and M spectra induced by duoplasmatron generated proton and oxygen ion beams, Nucl. Instr. Meth. Phys. Res. B (2017) (submitted for publication). [12] R.D. Deslattes, E.G. Kessler, P. Indelicato, L. deBilly, E. Lindroth, J. Anton, X-ray transition energies: new approach to a comprehensive evaluation. Review of Modern, Physics 75 (35) (2003).