Nuclear Instruments and Methods in Physics Research B 150 (1999) 554±558
PESA as a complementary tool to PIXE at CTU Prague Josef Voltr *, Jaroslav Kr al, Zdenek Nejedl y Faculty of Nuclear Sciences and Physical Engineering, Czech Technical University, V Holesovi ck ach 2, 180 00 Praha 8, Czech Republic
Abstract Proton Elastic Scattering Analysis (PESA) is a simple convenient method for hydrogen analysis in thin samples. A Proton Induced X-ray Emission (PIXE) target chamber was equipped with a PIPS detector for detection of forwardscattered protons. One of the objectives was to perform PIXE and PESA analyses of air particulate targets simultaneously. Tests and calibrations were ful®lled mainly with Mylar foils 1.5±6.5 lm thick in the proton energy region between 1.35 and 2.3 MeV. The energy dependence of scattering cross section is dierent from the Rutherford formula. Comparison of PIXE/PESA analysis with the Guelph Scanning Proton Microprobe Laboratory at the University of Guelph, Canada on seven aerosol samples was carried out. The intercomparison results validated our PESA/PIXE quality assurance protocol. In addition, repeated measurements of Gelman Te¯oTM ®lters indicated a gradual increase of hydrogen content by 1 ng/cm2 per 1 lC/cm2 proton dose. Ó 1999 Elsevier Science B.V. All rights reserved. PACS: 25.40.C; 61.80; 82.70.R Keywords: PESA; Forward scattering; Hydrogen analysis; Te¯on irradiation; PIXE
1. Introduction The Ion Beam Laboratory at the Faculty of Nuclear Sciences and Physical Engineering of the Czech Technical University in Prague performs Proton Induced X-ray Emission (PIXE) analysis [1] of aerosol [2] and other types of samples [3]. The PIXE method, however, does not provide concentrations of light elements like hydrogen, carbon, or oxygen. Therefore, various complementary techniques are generally used in parallel to PIXE, e.g., Proton Induced c-ray Emission
* Corresponding author. Tel.: 420 2 21912214; fax: 420 2 6884818; e-mail:
[email protected]®.cvut.cz
(PIGE), Nuclear Reaction Analysis (NRA), or Proton Elastic Scattering Analysis (PESA). In case of aerosol samples, light element data provided by the complementary techniques contribute to the total elemental budget and the user can then better estimate the organic mass and, in turn, the total reconstructed aerosol mass [4]. In addition, the correlation of hydrogen and sulphur concentrations has an important role in aerosol source identi®cation ± as shown in Ref. [5], a high hydrogen±sulphur correlation and the molar ratio approximately 10:1 is characteristic for antropogenic aerosols. In many other ®elds, the hydrogen is a common contaminant element [6], for example, in semiconductor technology the hydrogen concentration in¯uences long-term and radiation stability of MOS structures [7].
0168-583X/99/$ ± see front matter Ó 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 5 8 3 X ( 9 8 ) 0 0 9 5 1 - 3
J. Voltr et al. / Nucl. Instr. and Meth. in Phys. Res. B 150 (1999) 554±558
2. PESA method PESA is a method for thin sample analysis using forward scattering protons. It is adopted especially for the determination of hydrogen. Several processes play an essential role in this analysis (Fig. 1). The primary particle penetrates the sample and loses a small part of its energy in electronic collisions. In case of a close interaction between the particle and a target nucleus, binary elastic scattering occurs and an essential energy is transferred from the projectile to the target nucleus. The relationship between the scattering angle and the energy transferred is described by the kinematic factor [8]. The cross section for proton scattering on heavy atoms can usually be described by the Rutherford formula. In the case of proton scattering on hydrogen atom, the Rutherford formula is not valid [9]. Two eects are responsible: the particles are identical and at energies above 1 MeV, the nuclear forces take part in the collision. After the interaction, both particles, the scattered and the recoiled one, lose another small part of their energies along their way through the target material outwards.
555
cannot be greater than p/2. If the scattering angle is close to zero, the energy transfer between both particles is small and the peaks corresponding to the protons scattered on nuclei with dierent masses are not resolved well. The resolution improves with increasing angle. When the analysing beam is perpendicular to the sample, and the scattering angle is close to p/2, the distance travelled by the scattered particle through the target would be very long. As a consequence, the particle energy would be distorted by straggling. Another geometry limitation was the mechanical arrangement of our PIXE chamber. Therefore, detection angle of 45o was found most appropriate. A Canberra SPP50-14-100 PIPS detector with 13 keV/Am241 resolution was used to detect the scattered and the recoiled particles. Two ion beam analyses can be performed simultaneously: PIXE and PESA. In case of aerosol samples, the ®rst one is more important and the parameters of analysis (energy, ion beam diameter and current) are to be selected mainly for the high quality PIXE results. Then, the appropriate count rate of the PIPS detector may be reduced suitably by the use of a
3. Experimental set-up The arrangement of our PIXE/PESA target chamber is shown in Fig. 2. The choice of the scattering angle was an important task. Because the measured hydrogen atoms and the accelerated protons have the same mass, the scattering angle
Fig. 1. Principles of PESA analysis: Ion penetrates a thin sample, loses energy, then scatters at a target atom, and its energy is again decreasing on the way outwards.
Fig. 2. Arrangement of the PIXE/PESA target chamber.
556
J. Voltr et al. / Nucl. Instr. and Meth. in Phys. Res. B 150 (1999) 554±558
collimator in front of it. A collimator of 1 mm in diameter was used in these experiments. The electronics consisted of standard Canberra devices ± preampli®er 2003BT, ampli®er 2020 and ADC 8077. Both PESA and PIXE spectra were collected by a computer controlled multichannel analyser with multi-ADC option. The data were transported via LAN to further processing. The area of the hydrogen peak with the background subtracted is a measure of the hydrogen content.
removed from the calibration process. The results showed that the reproducibility of calibration was in the range of 8%. The energy dependence of calibration is shown in Fig. 3. Tests were also performed at slightly dierent geometries ± the samples were placed 0.5 and 1 mm o their regular position in the beam direction in order to simulate small dierences in ®lter positions. The small variations of sample position did not in¯uence the analytical results.
4. Calibration
5. Aerosol samples
The objective of calibration was to obtain the proportionality coecient between the hydrogen content and the yield of the scattered protons, and to validate the reproducibility and accuracy of our PESA set-up. We have de®ned the calibration coecient R
After the calibration, we have proceeded with the analysis of real samples. The initial quality control tests included an interlaboratory comparison with another PIXE lab using PESA. The Guelph Scanning Proton Microprobe (GSPM) laboratory was selected as the comparison partner due to its aerosol monitoring program GAViM [4]. The comparison included two groups of air particulate targets: indoor and outdoor samples. In both cases, the Gelman Te¯oTM [10] ®lters were used as a ®ltration medium. The ®rst group contained four samples of ®ne indoor aerosols collected at the CTU campus in Prague and was labelled as HAT 1±4. The second group of the intercomparison samples included GAViM network samples containing PM2.5 air particulate from a rural location. The samples were analysed
R
E
N ; cQ
where N is the area of hydrogen peak in the spectrum, c is the arial concentration of hydrogen in the sample (lg/cm2 ), and Q is the dose ± total integrated beam current through the target. Then, R includes the scattering cross section and the experimental geometry. It varies with the energy of the beam in same way as the cross section, and is approximately independent of the diameter of the beam spot. The calibration procedure was done with a set of thin foils. The reference samples were prepared from Mylar foils 1.5±6.5 lm thick. The foils were weighted by Mettler MT 5 microbalance with a charge neutralisation by alpha source at the University of Guelph, and the areal concentrations of hydrogen were then calculated. These foils were mounted in the same frames as are the aerosol samples. The system was calibrated at several proton energies in the 1.35±2.3 MeV range. The analytical conditions were kept similar to those for PIXE aerosol analysis: proton beam of 6 mm in diameter, ion currents 4±10 nA; and the proton doses for sucient spectrum statistics were 0.5±2 lC. Some samples showed systematic dierences, likely due to mechanical irregularities, and were
Fig. 3. Energy calibration of the hydrogen detection system. Calibration coecient R in counts/(lg/cm2 )/lC based on eighty individual measurements. Data were ®tted by the second order polynomial.
J. Voltr et al. / Nucl. Instr. and Meth. in Phys. Res. B 150 (1999) 554±558
in Guelph, transported to Prague, and then remeasured at the CTU in Prague. Te¯oTM ®lters become very fragile after irradiation and some of them break up in the process, therefore only results from three undamaged GAViM samples are presented. Some samples were analysed twice or 3times to estimate the spread of measured values. In the PESA spectra from Guelph, the area of hydrogen peak was obtained by a simple linear background subtraction. In Prague, where the background was higher, a special software was developed. The program allows user to set interactively the ®tting regions in the vicinity of the hydrogen peak, calculates background as a second-order polynomial, and subtracts it from the peak integral. The PIXE spectra were processed by GUPIX [11] software in both laboratories. The results of PESA for hydrogen, and of PIXE for selected elements are listed in Table 1. A more detailed discussion of PIXE analysis is given in Ref. [12], where the data of dominant elements agree within 10%. The PESA data vary up to 17%. However, the estimated spread of results obtained by repeated analysis of the same sample in both laboratories was also in the region of 10± 15%. The comparison of that limited series indicated a good quality of simultaneous PIXE/PESA analyses.
557
in Te¯on ®lter. To clarify this phenomenon, we have performed several tests. A blank ®lter was repeatedly analysed at the same spot and the dependence of hydrogen content on the proton dose was investigated. Fig. 4 shows these values together with a linear ®t. The experimental result indicates that the proton irradiation of Te¯oTM with a dose of 1 lC/cm2 increases the concentration of hydrogen approximately by 1 ng/cm2 . This value was determined for the proton beam energy of 2 MeV. When the experiment was repeated at 2.25 MeV proton energy, the hydrogen increase was by about 10% higher. A simple calculation indicates that every proton of the beam penetrating the sample increases the hydrogen content in the ®lter by about 100 atoms.
6. Te¯on irradiation eect Fig. 4. Dependence of the hydrogen concentration in ng/cm2 on the ion dose ± experimental points and a linear ®t. Proton beam energy 2250 keV.
During the PESA calibration measurements, we observed a small and growing amount of hydrogen
Table 1 Comparison of the elemental areal concentrations on Te¯on ®lters measured in Prague and Guelph H
CARE 382 CARE 383 CARE 389 HAT 1 HAT 2 HAT 3 HAT 4
Si
S
AVG
DIF (%)
AVG
DIF (%)
AVG
3726 2489 1130 562 678 1397 455
ÿ17.5 ÿ13.6 ÿ7.0 5.4 17.1 1.9 ÿ11.0
144.8 275.7 30.5 373.7 380.5 456.8 180.0
7.3 9.4 1.0 6.1 11.3 1.0 20.3
5205.5 4805.5 3153.3 361.4 459.2 1501.0 464.8
Zn DIF (%) 7.9 3.2 ÿ0.9 1.5 ÿ1.2 0.2 11.3
AVG
DIF (%)
63.9 36.5 30.7 49.8 47.0 142.3 58.4
ÿ0.8 ÿ2.7 ÿ10.8 ÿ9.8 ÿ2.5 ÿ4.3 0.9
AVG is the average value (Guelph+Prague)/2 in ng/cm2 , DIF denotes relative dierence (GuelphÿPrague)/AVG.
558
J. Voltr et al. / Nucl. Instr. and Meth. in Phys. Res. B 150 (1999) 554±558
The absolute values of the hydrogen concentration increase presented in the previous paragraph may include signi®cantly high errors ± about 15±20%; mostly due to poor count statistics at these low hydrogen amounts. The accuracy of the ion dose measurement in our system is 1±2%. These measurement were done at a pressure lower than 2 ´ 10ÿ3 Pa. Because the target chamber is equipped with a turbo molecular pump, the amount of hydrogen in the residual gas is relatively high. Therefore, a likely cause of this phenomenon is that C±F bonds are broken by ion irradiation and the ¯uorine atoms are substituted by hydrogen from the residual atmosphere. When the irradiated sample was exposed in open air, no remarkable change in hydrogen content was observed in the consequent analysis. 7. Conclusions A PIXE analytical setup was completed with a detection system for forward-scattered particles. A calibration of PESA for hydrogen analysis was performed. The comparison of PIXE/PESA analysis of limited series of aerosol samples executed jointly with the Proton Microprobe Laboratory in Guelph showed a good agreement of results. As a surprising phenomenon it was found that the amount of hydrogen content in Gelman Te¯oTM ®lters increases with proton irradiation.
Acknowledgements The authors acknowledge Mr. Z. Skutina, who manufactured the necessary mechanical parts of the experimental arrangement. This work was supported by CTU grant IG 3097491. References [1] S.A.E. Johansson, J. Campbell, K.G. Malmqvist, Particle Induced X-Ray Emission Spectrometry (PIXE), Willey, New York, 1995. [2] Z. Nejedl y, J. Kr al, J. Voltr, R. Krejcõ, E. Swietlicki, y, P. Kubõk, J. Svejda, J. Cern Nucl. Instr. and Meth. B 136±138 (1998) 981. [3] J. Kr al, J. Voltr, Z. Nejedl y, Nucl. Instr. and Meth. B 109/ 110 (1996) 167. [4] Z. Nejedl y, J.L. Campbell, W.J. Teesdale, C. Gielen, Nucl. Instr. and Meth. B 132 (1997) 489. [5] T.A. Cahill, R.A. Eldred, D. Wallace, B.H. Kusko, Nucl. Instr. and Meth. B 22 (1987) 296. [6] W.A. Lanford, Nucl. Instr. and Meth. B 66 (1992) 65. [7] M.A. Briere, F. Wulf, D. Brunig, Nucl. Instr. and Meth. B 45 (1990) 45. [8] Chu Wei-Kan, Backscattering Spectrometry, Academic Press, New York, 1978. [9] G. Deconning, Introduction to Radioanalytical Physics, Akademiai Kiad o, Budapest, 1978. [10] Te¯oTM membrane- stretched PTFE with PMP support ring, 25 mm in diameter, 3 lm pores, P/N: R2PI025, Gelman Sciences, Northampton, UK. [11] J.A. Maxwell, W.J. Teesdale, J.L. Campbell, Nucl. Instr. and Meth. B 95 (1995) 407. [12] Z. Nejedl y, PIXE analysis of environmental samples, Ph.D. thesis, CTU Prague, 1998.
