劉合凡 葛良全 謝希成 趙劍錕 羅耀耀
基于蒙特卡羅方法的XRF探測器立體角分析
劉合凡 葛良全 謝希成 趙劍錕 羅耀耀
(成都理工大學(xué) 地學(xué)核技術(shù)四川省重點(diǎn)實(shí)驗(yàn)室 成都 610059)
為研究探測器立體角在X射線熒光(X-ray fluorescence, XRF)分析儀的設(shè)計(jì)對(duì)測量結(jié)果的影響,運(yùn)用蒙特卡羅方法對(duì)XRF儀進(jìn)行建模。結(jié)果表明,探測器脈沖計(jì)數(shù)隨著探測器與樣品之間距離的增大指數(shù)衰減;隨著探測器立體角的增大,特征峰計(jì)數(shù)非線性增大,源峰探測效率指數(shù)遞減;探測器本征探測效率與探測器立體角無關(guān)。本文研究方法和結(jié)論可為一些XRF儀的設(shè)計(jì)提供參考。
探測器立體角,X熒光分析,蒙特卡羅方法,擬合優(yōu)度
在元素分析領(lǐng)域,X射線熒光(X-ray fluorescence, XRF)分析儀是最重要的分析儀器類型之一,廣泛應(yīng)用于采礦、環(huán)保、石化、材料、食品等領(lǐng)域[1–5]。XRF測量中,儀器計(jì)數(shù)率不僅與待測試樣本身的物理化學(xué)性質(zhì)有關(guān),還與激發(fā)源、試樣、探測器的相對(duì)幾何位置有關(guān)[6–8]。探測器的安放位置不同,所形成的探測器立體角不同,從而使儀器探測效率也不一樣。在可接受誤差范圍內(nèi),應(yīng)盡量提高儀器熒光計(jì)數(shù)和探測效率,進(jìn)而提高分析精度和縮短分析時(shí)間。而經(jīng)大量學(xué)者研究,美國洛斯阿拉莫斯實(shí)驗(yàn)室的通用大型蒙特卡羅程序MCNP軟件已經(jīng)成為核技術(shù)領(lǐng)域譜數(shù)據(jù)模擬和分析的重要軟件之一[9–13]。該軟件的模擬數(shù)據(jù)在其相對(duì)誤差小于5%的前提下,運(yùn)用建模模擬計(jì)算方法預(yù)測譜線強(qiáng)度的準(zhǔn)確度約達(dá)95%,而用該方法模擬未知成份樣品,其預(yù)測準(zhǔn)確度也能達(dá)90%–97%[14]。本文運(yùn)用MCNP軟件參考實(shí)際熒光分析儀的具體情況進(jìn)行建模,通過改變探測器立體角得到模擬計(jì)算結(jié)果,以觀察探測器立體角對(duì)XRF測量結(jié)果的影響情況。
1.1 探測器立體角的計(jì)算
設(shè)α為激發(fā)源光子在樣品o點(diǎn)處激發(fā)待測元素特征X射線時(shí)的入射角,β為樣品o點(diǎn)處產(chǎn)生待測元素特征X射線對(duì)應(yīng)的出射角。顯然激發(fā)源出射射線在樣品中的激發(fā)位置不同,α、β的大小也不一樣,如圖1所示。
圖1 X射線熒光測量中立體角關(guān)系推導(dǎo)示意圖Fig.1 Schematic diagram of deriving the desired X-ray collection angle in XRF measurement.
式中,dΩ單位為sr(球面度)。
于是:
將式(3)代入式(1),則探測器相對(duì)于x點(diǎn)的立體角為:
1.2 擬合優(yōu)度評(píng)價(jià)
面對(duì)大量的數(shù)據(jù)分析工作,往往需要通過一定的方程擬合才能夠發(fā)現(xiàn)數(shù)據(jù)之間的確切關(guān)系。而方程擬合的好壞直接揭示了數(shù)據(jù)之間內(nèi)在關(guān)系的準(zhǔn)確性,一般情況下用“擬合優(yōu)度”這一物理量來表征曲線方程對(duì)數(shù)據(jù)擬合程度的好壞。度量擬合優(yōu)度的統(tǒng)計(jì)量用“確定系數(shù)”R2表示,計(jì)算公式為:
式中,n表示原始數(shù)據(jù)個(gè)數(shù);yi表示第i個(gè)原始數(shù)據(jù);表示對(duì)應(yīng)的第i個(gè)擬合數(shù)據(jù);表示原始數(shù)據(jù)均值。粗略地講,擬合值R2越高,則說明方程對(duì)離散點(diǎn)擬合得越好;R2值越低,方程對(duì)離散點(diǎn)擬合得越差[15]。從式(5)可以看出,R2取值在0–1之間。R2越接近于1,說明模型選擇和擬合越好,數(shù)據(jù)預(yù)測也越成功。特別對(duì)于線性擬合,R2越大,則回歸方程擬合數(shù)據(jù)越好,線性關(guān)系越強(qiáng)。
本文模型以待測樣品下平面正中心為原點(diǎn),平行樣品建立Y坐標(biāo),垂直樣品建立Z坐標(biāo),如圖2所示。
激發(fā)源真實(shí)模擬美國Moxtek公司TUB00083系列X光管,采用半徑為0.04cm的面源,光子在面源內(nèi)進(jìn)行高斯分布型均勻抽樣;出射粒子方向與?Y方向夾角45°,散射角度為±23°,粒子抽樣參考方向與?Y方向夾角45°,抽樣概率為0:1;抽樣粒子能量為0.050MeV。探測器真實(shí)模擬美國Moxtek公司XPIN-XT系列探測器,部分結(jié)構(gòu)做了適當(dāng)簡化。探測器中軸線與樣品+Y方向夾角45°,探測器晶體為正圓柱,晶體橫截面積為13mm2,厚度為625μm;Be窗厚度為8μm,保護(hù)殼材料為純鋁。探測器晶體鍍有一層二氧化硅保護(hù)膜,膜厚度為1μm;鈹窗與保護(hù)膜之間填充空氣。待測樣品為純?cè)貥悠?,粒度理想,分布均勻,尺寸?5cm×1cm。模型截?cái)嗲騼?nèi)填充空氣,截?cái)嗲蛲庥谜婵仗畛?。模型?jì)數(shù)采用F8數(shù)據(jù)卡,記錄探測器中脈沖數(shù)的能量分布,計(jì)數(shù)對(duì)象為探測器晶體部分。能量箱設(shè)置有0能量箱(空箱)、Epsilon能量箱以及常規(guī)能量箱??障溆靡杂涗洸东@否定的、非模擬的電子碰撞記錄;Epsilon能量箱用來記錄穿過探測器而未被探測器記錄的粒子數(shù);常規(guī)能量箱能量范圍為1.01×10?5–0.050MeV。能量之間進(jìn)行線性插值,共記錄1024道能量計(jì)數(shù),模擬探測器響應(yīng)的每一道道址能量均與實(shí)驗(yàn)儀器保持一致。設(shè)置激發(fā)源粒子抽樣個(gè)數(shù)為2×108個(gè),模擬相對(duì)誤差小于2%。
圖2 蒙特卡羅方法模擬X射線熒光測量系統(tǒng)空間幾何布置圖Fig.2 Geometric layout for XRF measurement system by Monte Carlo method.
對(duì)于一個(gè)固定幾何位置的已知探測器,其最大立體角應(yīng)當(dāng)為定值。本文通過改變探測器晶體軸心位置(沿Z軸正方向,如圖2所示)產(chǎn)生相對(duì)于入射原級(jí)X射線對(duì)樣品激發(fā)點(diǎn)x處的探測器立體角變化,從而研究不同探測器立體角引起待測元素特征峰計(jì)數(shù)的變化情況。除H值變化外,探測器其他參數(shù)均保持不變。
3.1 特征峰計(jì)數(shù)與“探測器-樣品”距離H的關(guān)系
在激發(fā)源和樣品相對(duì)位置保持不變的情況下,探測器脈沖計(jì)數(shù)隨著探測器與樣品之間距離H的增大呈指數(shù)衰減,如圖3所示。隨著激發(fā)源與樣品之間距離H的逐漸加大,激發(fā)源入射原級(jí)X射線在樣品表面的照射量變小,激發(fā)樣品待測元素所產(chǎn)生特征X射線的概率也在變小,從而導(dǎo)致了探測器脈沖計(jì)數(shù)的減小。
圖3 蒙特卡羅方法模擬Cu樣品特征峰計(jì)數(shù)與探測器軸心距離的關(guān)系Fig.3 Relationship between Cu's X-ray characteristic fluorescence peak counts and ‘detector axis to specimen' distance by Monte Carlo calculation.
由圖3可以看出,當(dāng)H值大于20cm時(shí),探測器脈沖計(jì)數(shù)基本降為0。分析原因,除了空氣對(duì)射線的有限吸收作用外,在H值大于20cm時(shí)出射特征X射線幾乎已處于探測器立體角之外,即探測器已基本探測不到出射特征X射線,故特征峰計(jì)數(shù)幾乎為0。
3.2 特征峰計(jì)數(shù)與探測器立體角的關(guān)系
以式(4)計(jì)算出探測器立體角并將其作為橫坐標(biāo),得出特征峰計(jì)數(shù)與探測器立體角的關(guān)系,如圖4所示。
圖4 蒙特卡羅方法模擬Cu樣品特征峰計(jì)數(shù)與探測器立體角的關(guān)系Fig.4 Relationship between Cu's X-ray characteristic fluorescence peak counts and the desired X-ray collection angle by Monte Carlo calculation.
圖4 中,隨探測器立體角的逐漸增大,特征峰計(jì)數(shù)非線性增大,對(duì)CuKα脈沖計(jì)數(shù)進(jìn)行指數(shù)擬合,擬合優(yōu)度為0.99966。
3.3 探測效率與探測器立體角的關(guān)系
根據(jù)峰總比和源峰探測效率的計(jì)算公式[16],得出模型探測器峰總比和源峰探測效率數(shù)值,并以探測器立體角作為橫坐標(biāo),得到關(guān)系如圖5所示。
圖5 蒙特卡羅方法模擬Cu樣品峰總比、源峰探測效率與探測器立體角的關(guān)系Fig.5 Relationship of the Cu's ‘peak to source' ratio, ‘peak to total' ratio and the desired X-ray collection angle by Monte Carlo calculation.
由圖5可以看出,隨著探測器立體角的增大,源峰探測效率成指數(shù)遞減,而峰總比卻基本保持不變,維持在76.96%左右。這一結(jié)果與張明等[17]對(duì)NaI(Tl)探測器探測效率的模擬結(jié)果一致。
隨著探測器立體角變小,探測器沉積X射線的幾率變小,全能峰計(jì)數(shù)必然降低,故源峰探測效率變小。但由于立體角變小的同時(shí),全譜計(jì)數(shù)也跟著變小,故峰總比不一定會(huì)變小。圖5顯示了峰總比并不隨探測器立體角而變化。根據(jù)峰總比定義,探測器本征探測效率也必定保持恒定值,即探測器本征探測效率與探測器立體角無關(guān)。
本文根據(jù)具體的XRF儀進(jìn)行蒙特卡羅建模,分析了探測器立體角與特征峰計(jì)數(shù)和探測效率的關(guān)系,得出結(jié)論如下:
(1) 探測器脈沖計(jì)數(shù)隨著探測器與樣品之間距離的增大指數(shù)衰減;
(2) 隨著探測器立體角的增大,特征峰計(jì)數(shù)非線性增大;
(3) 隨著探測器立體角的增大,源峰探測效率指數(shù)遞減;
(4) 探測器本征探測效率與探測器立體角無關(guān)。
1 李建勝, 王仁波, 周蓉生, 等. 高靈敏度野外X熒光分析系統(tǒng)及其在銅礦勘探中的應(yīng)用[J]. 核技術(shù), 2002, 25(3): 211–218
LI Jiansheng, WANG Renbo, ZHOU Rongsheng, et al. A fieldwork high-sensitivity X-ray fluorescence analysis system and its application on exploration of copper mine[J]. Nuclear Techniques, 2002, 25(3): 211–218
2 唐堯基, 樊靜, 馮素玲. 環(huán)境水樣中痕量肼的熒光分析[J]. 分析試驗(yàn)室, 2003, 22(1): 48–52
TANG Yaoji, FAN Jing, FENG Suling. Fluorimetric method for determination of trace hydrazine in environmental water samples[J]. Chinese Journal of Analysis Laboratory, 2003, 22(1): 48–52
3 管嵩, 丁仕兵, 王越, 等. X熒光光譜法測量高鈦渣中主次元素[J]. 分析測試學(xué)報(bào), 2012, 31(Z1): 198–202
GUAN Song, DING Shibing, WANG Yue, et al. Determination of major and minor components in high-titanium slag by X-ray fluorescence spectrometry[J]. Journal of Instrumental Analysis, 2012, 31(Z1): 198–202
4 任翔, 葛良全, 張慶賢, 等. 原位X射線熒光測井井液的影響與校正[J]. 核技術(shù), 2009, 32(10): 756–760
REN Xiang, GE Liangquan, ZHANG Qingxian, et al. A method to correct effect of drilling fluid on in situ X-ray fluorescence logging[J]. Nuclear Techniques, 2009, 32(10): 756–760
5 高華娜, 趙海英, 王志宙, 等. 螺旋藻的PXRD和XRF分析[J]. 食品科技, 2008, 33(12): 270–273
GAO Huana, ZHAO Haiying, WANG Zhizhou, et al. Analysis on the spirulinas by PXRD and XRF[J]. Food Science and Technology, 2008, 33(12): 270–273
6 Mc Neillfe C. Optimization of measurement distance of109Cd K XRF system for obese subjects[J]. Nuclear Science and Techniques, 2003, 14(1): 86–89
7 劉艷芳, 賴萬昌, 謝希成, 等. 能量色散型X熒光分析儀光管、樣品、探測器距離的蒙特卡羅優(yōu)化[J]. 核電子學(xué)與探測技術(shù), 2011, 31(9): 1038–1042
LIU Yanfang, LAI Wanchang, XIE Xicheng, et al. Optimization of the distance among the detector, tube and samples in the EDXRF analyzer by the Monte Carlo method[J]. Nuclear Electronics & Detection Technology, 2011, 31(9): 1038–1042
8 Dryak P, Kovar P. Experimental and MC determination of HPGe detector efficiency in the 40–2754 keV energy range for measuring point source geometry with the source-to-detector distance of 25 cm[J]. Applied Radiation and Isotopes, 2006, 64(23): 1346–1349
9 San Miguel E G, Perez-Moreno J P, Bolivar J P, et al. A semi-empirical approach for determination of low-energy gamma-emmiters in sediment samples with coaxial Ge-detectors[J]. Applied Radiation and Isotopes, 2004, 61(2–3): 361–366
10 He J, Zhang C M. The accurate calculation of the Fourier transform of the pure Voigt function[J]. Journal of Optics A: Pure and Applied Optics, 2005, 7(10): 613–616
11 Gardner R P, Sood A, Wang Y Y, et al. Single peak versus library least-squares analysis methods for the PGNAA analysis of vitrified waste[J]. Applied Radiation and Isotopes, 1997, 48(10–12): 1331–1335
12 陸安詳, 王紀(jì)華, 潘立剛, 等. 便攜式X射線熒光光譜測定土壤中Cr、Cu、Zn、Pb和As的研究[J]. 光譜學(xué)與光譜分析, 2010, 30(10): 2848–2852
LU Anxiang, WANG Jihua, PAN Ligang, et al. Determination of Cr, Cu, Zn, Pb and As in soil by field portable X-ray fluorescence spectrometry[J]. Spectroscopy and Spectral Analysis, 2010, 30(10): 2848–2852
13 Sharshar T, Elnomer T. Efficient calibration of HPGe detectors for volume-source geometries[J]. Applied Radiation and Isotopes, 1997, 48(5): 695–697
14 Vincze L, Janssens K, Adams F, et al. A general Monte Carlo simulation of EDXRF spectrometer II: polarized monochromatic radiation, homogeneous samples[J]. Spectrochimica Acta, 1993, B50(3): 127–147
15 趙松山. 對(duì)擬合優(yōu)度R2的影響因素分析與評(píng)價(jià)[J]. 東北財(cái)經(jīng)大學(xué)學(xué)報(bào), 2003, 3(27): 56–58
ZHAO Songshan. Analysis on factors affecting the goodness of fit R2[J]. Journal of Dongbei University of Finance and Economics, 2003, 3(27): 56–58
16 喬錄成, 白立新, 張一云, 等. 高純鍺γ譜儀對(duì)環(huán)境樣品探測效率的模擬計(jì)算[J]. 四川大學(xué)學(xué)報(bào)(自然科學(xué)版), 2003, 40(2): 301–305
QIAO Lucheng, BAI Lixin, ZHANG Yiyun, et al. Simulating calculation of the HPGe γ-spectrometer detection efficiency for environmental samples[J]. Journal of Sichuan University (Natural Science Edition), 2003, 40(2): 301–305
17 張明, 王仁仲, 楊元第, 等. 蒙特卡羅方法模擬激發(fā)源位置對(duì)井型NaI(Tl)晶體探測效率的影響[J]. 核電子學(xué)與探測技術(shù), 2012, 32(1): 50–54
ZHANG Ming, WANG Renzhong, YANG Yuandi, et al. Influence of the source position to the well type NaI(Tl) detection efficiency based on a Monte Carlo simulation[J]. Nuclear Electronics & Detection Technology, 2012, 32(1): 50–54
CLC TL8, O657.3
Calculation of desired X-ray collection angle on XRF analyzer designed by Monte Carlo method
LIU Hefan GE Liangquan XIE Xicheng ZHAO Jiankun LUO Yaoyao
(Chengdu University of Technology, Key Laboratory of Applied Nuclear Techniques in Geosciences, Chengdu 610059, China)
Background: The designing of the X-ray fluorescence (XRF) analyzer's geometric layouts need to be considered, such as ‘detector to specimen' distance, ‘detector to source' distance, ‘source to specimen' distance. The desired X-ray collection angle is one of the important factors of the detection performance. However, the experience geometric layouts have been unable to meet every XRF analyzer designing, because the performance of the excitation source or the detector is getting better, sample processing technology is much more advanced, and so on. Purpose: The aim is to study the impact of the desired X-ray collection angle on XRF analyzer designing, and provide a technical guidance on methodologies for XRF analyzer designing. Methods: In this paper, we build the XRF analyzer models by the Monte Carlo method and analyze the impacts of the desired X-ray collection angle on XRF analyzer designing. Results: Kinds of factors with the desired X-ray collection angle are analyzed, such as Cu's X-ray characteristic fluorescence peak counts, the ‘detector axis to specimen' distance, the Cu's ‘peak to source' ratio. Conclusions: With the increasing of distance between the detector and the specimen, the detector's pulse counts satisfy an exponential decay law. With the desired X-ray collection angle increasing, the Cu's X-ray characteristic fluorescence peak counts increase linearly. With the desired X-ray collection angle increasing, the ‘peak to source' ratio decays exponentially, but the ‘peak to total' ratio remains the same.
Desired X-ray collection angle, XRF analysis, Monte Carlo method, Goodness of fit
TL8,O657.3
10.11889/j.0253-3219.2015.hjs.38.060502
國家863計(jì)劃項(xiàng)目(No.2012AA061803)、國家自然科學(xué)基金(No.41074093)資助
劉合凡,男,1983年出生,2010年于成都理工大學(xué)獲碩士學(xué)位,現(xiàn)為博士研究生,研究方向?yàn)楹诵畔@取與處理
葛良全,E-mail: glq@cdut.edu.cn
2014-09-22,
2014-11-20