CHIN. PHYS. LETT. Vol. 29, No. 11 (2012) 117901 Formula for the Probability of Secondary Electrons Passing over the Surface Barrier into a Vacuum * XIE Ai-Gen(谢爱根)** , XIAO Shao-Rong(肖韶荣), ZHAN Yu(詹煜), ZHAO Hao-Feng(赵浩峰) School of Physics and Optoelectronic Engineering, Nanjing University of Information Science and Technology, Nanjing 210044 (Received 5 April 2012) Based on a simple classical model that primary electrons at high electron energy interact with the electrons of lattice by the Coulomb force, we deduce the energy of secondary electrons. In addition, the number of secondary electrons in the direction of velocity of primary electrons per unit path length, 𝑛, is obtained. According to the energy band of the insulator, 𝑛, the definition of the probability 𝐵 of secondary electrons passing over the surface barrier of insulator into the vacuum and the assumption that lattice scattering is ignored, we deduce the expression of 𝐵 related to the width of the forbidden band (𝐸𝑔 ) and the electron affinity 𝜒. As a whole, the 𝐵 values calculated with the formula agree well with the experimental data. The calculated 𝐵 values lie between zero and unity and are discussed theoretically. Finally, we conclude that the deduced formula and the theory that explains the relationships among 𝐵, 𝜒 and 𝐸𝑔 are correct. PACS: 79.20.Hx DOI: 10.1088/0256-307X/29/11/117901 Secondary electron yield is an important and widely studied topic,[1−8] and the probability of secondary electrons passing over the surface barrier of the insulator into the vacuum, 𝐵, is an important research topic of secondary electron yield. Up to now, 𝐵 has been inaccessible to measurement. Alig and Bloom[9] have deduced the formula of 𝐵 related to the maximum secondary electron yield 𝛿𝑚 , the primary energy of 𝛿𝑚 , i.e. 𝐸𝑚 , and the average energy required to produce a secondary electron in an insulator, i.e. 𝜀; and a simple theoretical relationship of 𝐵 to 𝜒/𝐸𝑔 was developed to define an upper boundary for the 𝐵 calculated in Ref. [9], where 𝐸𝑔 and 𝜒 are the width of the forbidden band and electron affinity, respectively. In this Letter, based on a simple classical model that primary electrons at a high electron energy interact with the electrons of the lattice by the Coulomb force, the energy of secondary electrons is deduced. Moreover, the number of secondary electrons in the direction of velocity of primary electrons per unit path length (i.e. 𝑛) is obtained. According to the energy band of the insulator, 𝑛, the definition of 𝐵 and the assumption that lattice scattering is ignored, the formula for 𝐵 related to 𝐸𝑔 and 𝜒 is deduced, and the theory that explains the relationships among 𝐵, 𝜒 and 𝐸𝑔 is developed and proven to be correct. The formula for 𝐵 deduced here is universal for the estimation of 𝐵 under the condition that primary electrons at any energy hit on insulators. Primary electrons at high electron energy enter the emitter perpendicularly and travel along a straight line; the minimum distance between the electrons of the lattice and the primary electrons is 𝑠, and the time when primary electrons pass the minimum distance between the electrons of lattice and primary electrons is zero, as shown in Fig. 1, the distance 𝑟 between the electrons of the lattice and the primary electrons can be expressed as follows:[10] 𝑟 = (𝑠2 + 𝑣 2 𝑡2 )1/2 , (1) where 𝑣 is the velocity of primary electrons, and 𝑡 is the time measured from the instant of minimum distance. Based on the process of movement of primary electrons, the momentum of an electron of a lattice is perpendicular to the velocity of the primary electron and can be expressed as follows:[10] ∫︁∞ 𝑒2 𝑠𝑑𝑡 2𝑒2 ∆𝑝 = = , (2) 𝑣𝑠 (𝑠2 + 𝑣 2 𝑡2 )3/2 −∞ where 𝑒 is electronic charge. The energy of secondary electrons, 𝐸1 , can be written as[10] 𝑒4 1 (∆𝑝)2 = , (3) 𝐸1 = 2𝑚 𝐸𝑝 𝑠2 where 𝐸𝑝 = 𝑚𝑣 2 /2 is the primary energy, and 𝑚 is the electron mass. We draw a cylindrical shell in the direction of velocity of the primary electrons, the radius of the cylindrical shell is in the range [𝑠, 𝑠 + ∆𝑠], the number of secondary electrons produced in the cylindrical shell per unit path length can be written as ∆𝑛 = (2𝜋𝑠∆𝑠)𝑁 = 𝜋𝑁 ∆(𝑠2 ), (4) where 𝑁 is the number of secondary electrons per unit volume. Based on Eqs. (3) and (4), the number of secondary electrons produced in the cylindrical shell per unit path length can be written as * Supported ∆𝑛 = 𝜋𝑁 𝑒4 1 𝜋𝑁 𝑒4 ∆𝐸1 ∆( ) = . 𝐸𝑝 𝐸1 𝐸𝑝 𝐸12 by the Natural Science Foundation of Jiangsu Provincial Universities under Grant No 10KJB180004. author. Email: xagth@126.com © 2012 Chinese Physical Society and IOP Publishing Ltd ** Corresponding 117901-1 (5) CHIN. PHYS. LETT. Vol. 29, No. 11 (2012) 117901 The number of secondary electrons produced in the cylinder whose radius is 𝑠, per unit path length, can be written as ∫︁ ∑︁ 𝜋𝑁 𝑒4 𝑑𝐸1 . (6) 𝑛= ∆𝑛 = 𝐸𝑝 𝐸12 into the vacuum, produced in the cylinder whose radius is 𝑠, per unit path length, can be written as When primary electrons at a high electron energy hit on the insulator, most of the secondary electrons can be excited from the valence band. The role of the lattice is minor.[11] We assume that lattice scattering is ignored, therefore, according to the energy band of the insulator, as shown in Fig. 2, the minimum (i.e. 𝐸min ) is 𝐸𝑔 . The secondary electrons excited at distances larger than the effective escape depth can not be emitted into the vacuum,[14] and the probability of secondary electrons with energy 𝐸 > 𝜒 reaching the surface can be written as[15] 𝜋𝑁 𝑒4 ∆𝑛 = 𝐸𝑝 (𝐸𝑔 +𝜒+50) ∫︁ (𝐸𝑔 +𝜒) 𝑑𝐸1 . 𝐸12 𝑓 (𝑥) = exp(−𝛼𝑥), (9) (10) where 𝑥 is the distance from the surface of emitter, and 1/𝛼 is the effective escape depth of secondary electrons with energy 𝐸 > 𝜒. According to Eqs. (9) and (10), the number of secondary electrons which can reach the insulator surface barrier, produced in the distance range of 0 ≤ 𝑥 ≤ 1/𝛼, can be written as Electron of lattice r S νt 𝑛𝑝1 = ∑︁ Primary electron Fig. 1. Distance between the electron of the lattice and the primary electron. 𝑛𝑝2 ∫︁1/𝛼 = 𝑛𝑝1 exp(−𝛼𝑥)𝑑𝑥 0 Vacuum level (𝑒 − 1)𝜋𝑁 𝑒4 = 𝑒𝛼𝐸𝑝 χ Bottom of conduction band (𝐸𝑔 +𝜒+50) ∫︁ (𝐸𝑔 +𝜒) 𝑑𝐸1 , 𝐸12 (11) 𝐸1 can be written as Eg Top of valence band 𝐸1 = 𝐸 + 𝐸𝑔 , Fig. 2. The energy band of the insulator. Classically,[12] the electrons in the conduction band of the insulator with 𝐸 > 𝜒 can pass over the surface barrier into a vacuum, where 𝐸 is measured from the bottom of the conduction band. On the basis of the energy band of the insulator, the conservation of energy and the above assumption, the minimum average energy required to produce a secondary electron which can pass over the insulator surface barrier into vacuum can be written as 𝐸minipi = 𝜒 + 𝐸𝑔 . (12) where 𝐸1 is measured from the top of the valence band. The probability of a secondary electron with energy 𝐸 reaching the surface escapes from the solid is 𝑃 (𝐸), classically, 𝑃 (𝐸) = 0 for 𝐸 ≤ 𝜒. For 𝐸 > 𝜒, 𝑃 (𝐸) can be written as[12] 𝑃 (𝐸) = 1 − (𝜒/𝐸)1/2 . Based on Eqs. (11)–(13), electrons which can pass barrier into the vacuum, range of 0 ≤ 𝑥 ≤ 1/𝛼, can (𝑒 − 1)𝜋𝑁 𝑒4 𝑛𝑝 = 𝑒𝛼𝐸𝑝 (7) (13) the number of secondary over the insulator surface produced in the distance be written as (𝜒+50) ∫︁ [1 − (𝜒/𝐸)1/2 ] 𝑑𝐸. (𝐸𝑔 + 𝐸)2 (14) 𝜒 Secondary electrons are conventionally defined as having exit energies 𝐸𝑠 ≤ 50 eV while backscattered electrons have 𝐸𝑏 > 50 eV,[13] therefore, based on the energy band of the insulator, the conservation of energy, the definition of secondary electrons and the above assumption, the maximum 𝜀 can be written as 𝐸max = 𝜒 + 𝐸𝑔 + 50. Based on Eqs. (6)–(8), the number of secondary electrons produced in the cylinder whose radius is 𝑠 per unit path length can be written as 𝑛ti = (8) Based on Eq. (6)–(8), the number of secondary electrons which can pass over the insulator surface barrier ∑︁ 𝜋𝑁 𝑒4 ∆𝑛 = 𝐸𝑝 (𝐸𝑔 +𝑥+50) ∫︁ 𝐸𝑔 𝑑𝐸1 . 𝐸12 (15) The probability of secondary electrons with energy 𝐸 > 𝜒 reaching the surface can be written as Eq. (10), 117901-2 CHIN. PHYS. LETT. Vol. 29, No. 11 (2012) 117901 that is to say, only a part of the secondary electrons with energy 𝐸 > 𝜒, produced in the distance range of 0 ≤ 𝑥 ≤ 1/𝛼, can reach the surface. The probability of secondary electrons with energy 0 ≤ 𝐸 ≤ 𝜒 reaching the surface can be written as[15] 𝑓 (𝑥) = exp(−𝛼2 𝑥), (16) where 1/𝛼2 is the effective escape depth of secondary electrons with energy 0 ≤ 𝐸 ≤ 𝜒, the energy of secondary electrons with energy 0 ≤ 𝐸 ≤ 𝜒 is lower than the energy of secondary electrons with energy 𝐸 > 𝜒, so 1/𝛼2 is larger than 1/𝛼;[16] only a part of secondary electrons with energy 0 ≤ 𝐸 ≤ 𝜒, produced in the distance range of 0 ≤ 𝑥 ≤ 1/𝛼2 , can reach the surface. Thus we can think that the number of secondary electrons reaching the surface is approximately equal to the number of secondary electrons produced in the distance range of 0 ≤ 𝑥 ≤ 1/𝛼. Therefore, according to Eq. (15), the number of secondary electrons reaching the surface can be written as 𝑛𝑡 = 𝑛ti 𝛼 𝜋𝑁 𝑒4 = 𝛼𝐸𝑝 (𝐸𝑔 +𝑥+50) ∫︁ 𝐸𝑔 𝜋𝑁 𝑒4 (𝜒 + 50) 𝑑𝐸1 = . 𝐸12 𝛼𝐸𝑝 𝐸𝑔 (𝐸𝑔 + 𝜒 + 50) (17) On the basis of Eqs. (14) and (17), the formula for 𝐵 can be written as 𝑛𝑝 0.63𝐸𝑔 (𝐸𝑔 +𝜒+50) 𝐵= = 𝑛𝑡 (50+𝜒) 𝜒+50 ∫︁ 1 − (𝜒/𝐸)1/2 𝑑𝐸. (𝐸𝑔 +𝐸)2 𝜒 (18) Alig and Bloom[9] have given the formula for 𝐵, i.e. 𝐵 = 2.5𝛿𝑚 𝜀/𝐸𝑚 , (19) where 𝜀 is written as 𝜀 = 2.8𝐸𝑔 . (20) Seen from Table 1, the 𝐵1 calculated with Eq. (18) and the parameters in Table 1 agree well with the 𝐵2 calculated with Eqs. (19) and (20). However, for the insulators whose band gap is very narrow, such as InSb, Cs3 Bi, Ge and PbS, the 𝐵2 calculated with Eqs. (19) and (20) and the parameters in Table 1 are much fewer than the 𝐵1 calculated with Eq. (18) and with the parameters in Table 1, the reason for is that the 𝜀 calculated with Eq. (20) and the parameters in Table 1 are much fewer than the experimental values of the insulators whose band gap is very narrow, for example, the 𝜀 of InSb calculated with Eq. (20) and the parameters in Table 1 are equal to 0.56 eV and are much fewer than the experimental data equal to 1.2 eV.[17] For another example, the 𝜀 of PbS calculated with Eq. (20) and with the parameters in Table 1 is equal to 1.12 eV and is much smaller than the experimental values which are equal to 2.0 eV.[18] As shown in Table 1, the 𝐵2 of MgO calculated with Eqs. (19) and (20) and with the parameters in Table 1 is much larger than the 𝐵1 of MgO calculated with Eq. (18) and with the parameters in Table 1 it is greater than unity. Because it is impossible for 𝐵 to be larger than unity, we think that the 𝐵2 of MgO calculated with Eqs. (19) and (20) and with the parameters in Table 1 is not reasonable and that the 𝐵1 of MgO calculated with Eq. (18) and with the parameters in Table 1 is reasonable. Table 1. The calculated 𝐵1 and 𝐵2 . Insulator 𝐸𝑔 (eV)[9] 𝜒 (eV)[9] 𝐵1 𝛿𝑚 [9] 𝐸𝑚 (eV)[9] KF 10.9 0 0.63 11 1100 9.0 0.5 0.4 12 1700 NaCl 14 1200 11 1000 8.5 0.5 0.4 8 1500 KCl 13 1500 RbCl 8.3 0.5 0.4 11 1200 LiBr 8.0 0.2 0.47 8 1000 7.5 0.4 0.41 16 2200 NaBr 23 1600 7.6 0.8 0.34 10 1500 KBr 14 1500 13 1400 RbBr 7.2 0.4 0.41 11 1700 LiI 6.0 2.0 0.21 6 850 NaI 5.9 1.5 0.25 11 1100 11 1400 6.2 1.2 0.28 8 1600 KI 11 1500 RbI 5.8 1.2 0.27 10 1700 CsI 6.4 0.3 0.42 20 2600 7.8 0.9 0.33 24 1300 MgO 10 400 7.7 0.7 0.36 5 600 CaO 28 2600 BaO 5.1 0.6 0.34 4.7 450 ZnO 3.4 4.6 0.085 1.7 300 9.9 1.1 0.33 4 500 Al2 O3 6.4 650 SiO2 11 1.0 0.35 3 500 GaN 3.2 4.1 0.09 2.8 600 PbS 0.4 3.4 0.019 1.2 500 GaAs 1.4 4.0 0.049 1.1 350 InSb 0.2 4.6 0.007 1.15 700 C 5.5 6.3 0.087 2.8 750 Si 1.1 4.0 0.04 1.1 250 0.7 4.1 0.026 1.15 500 Ge 1.2 400 1.7 4.5 0.051 1.4 350 Se 1.4 400 Cs3 Sb 1.6 0.5 0.24 6 700 Rb3 Sb 1.0 1.2 0.11 7.1 450 Cs3 Bi 0.6 1.3 0.07 3.5 650 𝐵2 0.76 0.44 0.74 0.69 0.32 0.52 0.53 0.45 0.38 0.75 0.35 0.50 0.49 0.33 0.3 0.41 0.32 0.22 0.32 0.24 0.34 1.01 1.36 0.45 0.58 0.37 0.13 0.55 0.68 0.46 0.10 0.0067 0.031 0.0023 0.14 0.034 0.011 0.015 0.048 0.042 0.096 0.11 0.023 As seen from Table 1, the 𝐵1 values calculated with Eq. (18) and with the parameters in Table 1 lie between zero and unity and are consistent with predictions obtained using a free-particle approximation. It is concluded that Eq. (18) and the theory that explains 117901-3 CHIN. PHYS. LETT. Vol. 29, No. 11 (2012) 117901 the relationships among 𝐵, 𝜒 and 𝐸𝑔 are correct. For a secondary electron emitter, 𝐵 is a constant and is not a function of the incident energy of primary electrons.[1,2,4−6,15,19] Therefore, Eq. (18) for 𝐵 deduced under the condition that primary electrons at a high electron energy hit on insulators is universal for the estimation of 𝐵 under the condition that primary electrons at any energy hit on insulators. Based on a simple classical model that primary electrons at high electron energy interact with the electrons of lattice by the Coulomb force, the energy band of the insulator, the conservation of energy and the assumption that lattice scattering is ignored, we successfully obtain Eq. (18) for 𝐵 and it is universal for the estimation of 𝐵 under the condition that primary electrons at any energy hit on insulators. The theory that explains the relationships among 𝐸𝑔 , 𝜒 and 𝐵 is correct. References [1] Xie A G, Zhang J and Wang T B 2011 Chin. Phys. Lett. 28 097901 [2] Xie A G, Li C Q and Wang T B 2009 Mod. Phys. Lett. B 23 2331 [3] Lei W, Zhang X B and Zhou X D 2005 Appl. Surf. Sci. 251 170 [4] Xie A G, Zhao H F, Song B and Pei Y J 2009 Nucl. Instrum. Methods Phys. Res. Sect. B 267 1761 [5] Zhao S L and Bertrand P 2011 Chin. Phys. B 20 037901 [6] Xie A G, Zhang J and Wang T B 2011 Jpn. J. Appl. Phys. 50 126601 [7] Kazami Y, Junichiro K and Norio O 2009 Appl. Surf. Sci. 256 958 [8] Yasuda M, Tamura K and Kawata H 2001 Appl. Surf. Sci. 169 78 [9] Alig R C and Bloom S 1978 J. Appl. Phys. 49 3476 [10] Baroody E M 1950 Phys. Rev. 78 780 [11] Smith A C, Janak J F and Adler R B 1967 Electron. Conduction Solids (McGraw-Hill New York) P 235 [12] van der Ziel A 1968 Solid State Physical Electronics 2nd edn (Englewood Cliffs: Prentice-Hall) p 172 [13] Reimer L and Drescher H 1977 J. Phys. D 10 805 [14] Copeland P L 1940 Phys. Rev. 58 604 [15] Seiler H 1983 J. Appl. Phys. 54 R1 [16] Nishimura K, Kawata J and Ohya K 2000 Nucl. Instrum. Methods Phys. Res. Sect. B 164 903 [17] Tauc J and Abraham A 1959 Czech. J. Phys. 9 95 [18] Smith A and Dutton D 1958 J. Opt. Soc. Am. 48 1007 [19] Fijol J J, Then A M and Tasker G W 1991 Appl. Surf. Sci. 48 464 117901-4