Surface Plasmon Resonance (SPR) Theory: Tutorial

Masahiro Yamamoto
Department of Energy and Hydrocarbon Chemistry, Kyoto University, Kyoto-Daigaku-Katsura, Nishikyo-ku, 615-8510, JAPAN This manuscript is modied on October 20, 2008 8 : 22 am

Introduction

In the surface plasmon resonance (SPR) measurement we can detect the change of the lm thickness (refractive index change) in the sub-nonometer scale using a low-cost home-made appratus (He-Ne laser, rotating stage, photo-diode detector, prism, maching oil, and 50 nm gold lm on a glass plate, See Appendix). In this sense SPR is widely used, in particular, as the immunoassay to detect label-free antigen-antibody reaction. When light is irradiated to the optical prism |thin metallic lm (usually 50 nm gold lm is used)| sample system, the reectivity of the light becomes almost zero at the angle of incidence where the surface plasma wave of gold surface can couple to the part of the incident light. 1 This angle is called SPR angle and is very sensitive to the lm thickness (refractivity) of the sample. In the usual automated SPR measurementanalysis routine one can get the SPR angle shift and the the change of the lm thickness (and/or refractivity), thereby the measurement may be one of the black-box. However I think it is important to know the principle for the people to do something new by an detailed analysis of the measurements. From the name plasmon one may misunderstand that the quantum mechanical understanding is required for SPR principles, but the phenomena can be understood from the classical optics (or electromagnetic theory) which explain the light reection, transmission, and absorption for the multi-layer medium. In the famous book by Raether the principle of SPR was written completely, but at least for me it is not easy to understand because the book may be for specialist. 2 This note is written for my self-study of SPR principle in order to analyze the SPR curves, 3 . ,later was publish in Review of Polarography, 48, No3. 2002, 209 , and posted the revised version at http://www.scl.kyoto-u.ac.jp/ masahiro/sprtheory.html. (I am very happy that many people give me contacts and some comments.) In this note on SPR theory I also describe the basics of the light reection, transmission, absorption at the surface of dielectrics or metal, then it is my pleasure that

Plasma oscillation is the phenomena that free electrons in the metal oscillate cooperatively from their equilibrium position where the positive charge of metal (atomic nucleus or jellium that the positive charges are averaged) bind the ensemble of the free electrons . Plasmon means the quasi-particle representation of plasma frequency. The surface plasmon is the plasma oscillation that localize at the surface or interface. 2 There are a lot of review or papers which describe the SPR principle, but almost all of them did not describe the formulation and start from the equation of the SPR resonance condition. I want to formulate the equation from scratch. 3 Tell me, and I forget. Teach me, and I may remember. Involve me , and I learn. Benjamin Franklin

those who are doing the photon-in/photon-out electrochemistry with the interface spectroscopy with evanescent wave, ellipsometry, nonlinear spectroscopy, have an interest in this note. The SPR(surface plasmon resonance)[1] is the century-old technique from the nding of the Woods anomaly for the reected light from the diraction gratings[2]. After Ottos demonstration[3] for the surface plasmon excitation by light with attenuatedtotal-reection(ATR) coupler[4], the SPR method applied to the organic lms[5] or the detection of antigen-antibody reaction[6]. The SPR theory is also well established[7], and the recent advance in the measurements can be reported in the reviews[8, 9, 10]. In the SPR method the dielectric constant change in the sub-nm region from the surface can be measured and the method can be easily applied to the adsorption phenomena in the electrochemical environment[11, 12], where the capacitance can be measured simultaneously and get the complementary information of the change in the dielectric properties on the electrode surface. In this note Theoretical Electromagnetics(Japanese, Kinokuniya, 1973) by S. Sunagawa, Principles of Optics (7th expanded edition) (Cambridge Univ. Press, 1999) by M. Born and E. Wolf[13], and Optics(4th edition) (Addison Wesley, 2002) by E. Hecht[14] are used for reference.

Maxwell Equation
4

The Maxwell equations are described

divD = divB = 0 B t D rotH = J + t rotE =

(1) (2) (3) (4)

Here we use MKSA-SI unit. The electric eld E (Vm1 ) and magnetic eld H (Am1 ) are related to the electric displacement (or dielectric ux density or electric ux density) D (Cm2 ) and magnetic-ux density (or magnetic induction) B (T:tesla = NA1 m1 ) D = 0 E B = 0 H
4

(5) (6)

The Maxwell equations have the beautiful mathematical form. We learned the physical meaning of the equations in high school physics course and are mathematically extended to the 3-dimensional case (or 4 dimension when time is included.) by vector analysis. divV V means the dierence between the input of the vector eld V(r) to the innitesimal volume dr and the output from there, and is called as divergence. In the vector analysis the meaning of rot is hard to understand, but S. Naganuma gave us a method to understand the meaning of rotation in his book (Intuitive method for phycal mathematics(Japanese) ) [The rotational velocity of the innitesimal water wheel in water ow eld u. (The water ow of the right side of the wheel in the upper direction uy is faster than the left side (uy /x > 0), the wheel rotates in the anticlockwise direction. In the same the water ow of the upper side of the wheel in the right direction ux is slower than the lower side (ux /y < 0) the wheel rotates in the anticlockwise direction. (rotu)z = uy /x ux /y ) ] The Maxwell equations say, roughly speaking, Eq.(1): The number of the electric ux line coming(entering) from the positive (negative) charge is proportion to the amount of the charge. Eq.(2): magnet has always S and N pole. Eq.(3): When a magnet approachs the coil, the inductive current is passed to reduce the magnetic eld from the magnet. Eq.(4): A current generates the rotating magnetic eld in the direction of the right-handed screw.

Here and 0 are the dielectric constant (with no dimension) and electric permittivity of free space [8.854187817 1012 Fm1 (= CV1 m1 )], respectively. and 0 are magnetic permeability (with no dimension) and magnetic permeability of free space (4 107 NA2 ), respectively. We will assume the Ohms law for the relation between the current J and the electric eld E J = E (7)

2.1

Energy Conservation and Poynting Vector

5

The equation of motion of point charges is mi ri = If we apply (

i vi )

dr {ei (r ri (t))E + ei (r ri (t)) ri B}

i from the left, and the velocity is dened as vi = r

i = mi vi v
i i

dr ei (r ri (t))vi E + ei (r ri (t)) vi [vi B]

=0

=
i

drei (r ri (t))vi E

(8)

From the denition of the current and the Eq.(4). J =

i

i (t) (r ri (t)) ei r D t D t

(9) (10)

= rotH Then d dt 1 2 mi vi 2 =

dr rotH

(11)

D B 1 (E D + B H) = E +H 2 t t t D = E H rotE t 1 1 d 2 = dr mi vi (E D + B H) dt 2 2 t i

H rotE + E rotH
=div(EH)

d dt

1 2 + mi vi 2

1 2

dr(E D + B H)

dS

[E H]
Poynting vector

n (12)

kinetic energy

total energy of electromagnetic eld

From above equations the Poynting vector S[= E H] means the energy ux going out from the system.
5 The point charge, which is given by delta function, feel the Lorentz force from the electromagnetic eld.

2.2

Wave Equations
rotE = 0 H t (13)

From Eqs.(3) and (6)

From Eqs.(4), (5), and (7) E t If we apply to Eq.(13) and /t to Eq.(14) and using the relation rotH = E + 0 i E = rot
x

(14)

j
y

k
z

i =
Ez y x

j
y Ey z Ex z

k
z Ez x Ey x

Ex Ey Ez = i

Ex y

2 Ey 2 Ex 2 Ez 2 Ez + + j[..] + k[..] 2 2 xy y z xz Ex Ey Ez = i + + + j divE + k divE x x y z y z 2 2 2 i Ex j Ey k Ez = grad(divE) 2 E If we assume = 0, divE = 0 then 2 E = 0 In the same way we can get 2 H = 0 H 2H + 0 0 2 t t (17) 2E E + 0 0 2 t t (16) (15)

In the case that the electric eld has a plane wave form E = E0 ei(krt) (18)

where E0 is the polarization vector and the wavevector k is in the direction of the wave propagation and the magnitude is given from Eq.(16) k 2 = i0 + 0 0 2 (19)

In vacuum, = 1, = 1, = 0 and c/ = , c/ = /(2 ), k = 2/ = /c then c = 1/ 0 0 The complex optical index n may be given by k= n 2 = = v c 0 (21) (22) (23) (20)

n 2 = k 2 c2 / 2 = + i n = n + i,

k=n /c = (n + i)/c

where n and are the real and imaginary part of the complex optical index, respectively. E (z, t) = E0 ei(kz t) = E0 ez/c ei(nz/ct) I (z ) E E = |E0 | e = 2/c n
2 2 2z/c

(24)

= I0 e

(Lambert s Law) (25) (26) (27)

= = 1 + i2
2 2

[ = 1, = 0 in Eq.(22)]

1 = n ,

2 = 2n

Here is the absorption coecient, is the complex dielectric constant, 1 and 2 are the real and imaginary part of the complex dielectric constant, respectively. If we take divergence of the plane-wave electric led, divE = +i +i (E0x i + E0y j + E0z k)ei(kx x+ky y+kz z t) x y z = i(kx E0x + ky E0y + kz E0z ) = ik E0 = =0 0 i (28) (29) (30)

Then we can nd that the electric eld is transverse wave, i.e. k E0 . If the magnetic-ux density B is written as B= k E0 i(krt) e , (31)

then the magnetic-ux density satises Eq.(3), because rotE = i( Ez Ey ) + .. y z = i(iky Ez ikz Ey ) + .. (32) (33) (34) (35) (36)

B t

= ik E k E0 = (i )ei(krt) = ik E 0 = 376.7 0

In vacuum

|E0 | 0 0 = = = |H0 | k 0 0

(37)

2.3

Boundary Conditions at a Interface Between Dierent Media

Now we think a interface which is at the boundary medium 1 and 2 as shown in Fig.1. From Gauss law, we can get the following for the Gauss box which include the interface inside the box,
V

dr D =

D dS

(38)

In the limit that the Gauss box is very thin (h 0) dr = Qbox = n [D1 D2 ]S (39)

where vector n means the surface normal unit vector pointing from media 2 to 1. n [D1 D2 ] = Qbox /S = 12 5 (40)

where 12 means the interface charge density. From B = 0 n [B1 B2 ] = 0 As shown in Fig. 2 for a general vector eld V(r), Stokes theorem gives V dr = (I) + (II) + (III) + (IV) = Vx (x0 , y0 )dx + Vy (x0 + dx, y0 )dy + Vx (x0 , y0 + dy )(dx) + Vy (x0 , y0 )(dy ) Vy = Vx (x0 , y0 )dx + [Vy (x0 , y0 ) + dx]dy x Vx [Vx (x0 , y0 ) + dy ]dx Vy (x0 , y0 )dy y Vy Vx = dxdy = (rotV)z dxdy x y (41)

medium 1 n

r t1 h

medium 2

t2

Figure 1: Gauss and Stokes box

y (x0, y0+dy) (III) (x0+dx, y0+dy)

(IV) (I) (x0, y0)

(II)

(x0+dx, y0) x

Figure 2: Stokes theorem From Eq.(3) and Stokes theorem we can get
C

E dr =

rotE dS

(42) (43) (44) (45) (46)

B bdS S t B r[E1 t1 + E2 t2 ] = brh 0 t t = t1 = t2 = t [E1 E2 ] = 0 6

Here b is the unit vector dened by b = n t. From Eq.(4) H dr = = rotH dS J+ (47) (48) (49) (50)

D bdS t S D r[H1 t1 + H2 t2 ] = J+ brh J brh t t [H1 H2 ] = Js [ surface current density(Am1 )]

Reection and Transmission

k E0 i(krt) e k R0 i(k r t) e k T0 i(k r t) e

Now we dene the incident plane wave as E = E0 ei(krt) , reected wave as E = R0 ei(k r t) , transmitted wave as E = T0 ei(k
r t)

= (k ),

B=

(51)

= (k ),

B =

(52)

= (k ),

B =

(53)

Ep Es k Rs
x z

Rp k medium 1 x medium 2 Ts k Tp

Figure 3: Reection and Transmission of Light

From Fig.3 k = (k sin , 0, k cos ) k k

(54)

= (k sin , 0, k cos ), = (k sin , 0, k cos )

i( )

= cos + i sin

(55) (56)
6

The p-wave has the components of x and z , but s-wave has only the y component.

3.1

Condition I:

t [E1 E2 ] = 0
Ex + Ex = Ex

From Eq.(46) the tangential component of the electric eld become at z = 0

i(k sin x t)

(57) (58)

Ep cos e

i(k sin xt)

+ Rp cos e

i(k

sin x t)

= Tp cos e

For any x at z = 0 this condition should be satised, then = = k sin = k sin = k sin The magnitude of the wavevector is given by Eq.(21) sin = sin = sin( ), because k = k (61)

(59) (60)

The incident angle equals to the reection angle, and k sin = n 1 sin = k sin = n 2 sin c c (62)

Then we can get Snells law because = n 1 sin = n 2 sin The Eq.(58) becomes From Fig.3 we can get Ep cos Rp cos Tp cos Es Rs Ts E0 = , R0 = , T0 = Ep sin Rp sin Tp sin For y -direction we can get the condition
Ey + Ey = Ey

(63)

(Ep Rp ) cos = Tp cos

(64)

(65)

at z = 0
i(k sin xt)

(66) (67) (68)

(Es + Rs )ei(k sin xt) = Ts e (Es + Rs ) = Ts

6

Light (electromagnetic) wave is the transverse wave and has the two polarization components. (If you look at the the digital watch(almost linear polarized light!?) through the polarized sunglasses, you can see the bright and dark light at every 90 degree if you rotate the watch.) s-wave (s polarized light) is the component parallel to the reection plane, and p-wave (p polarized light) is the component perpendicular to the paper plane. Because the electric component is perpendicular to the paper plane, s-wave is called as TE (transverse electric) wave. Because the magnetic component is perpendicular to the paper plane, p-wave is called as TM wave.

3.2

Condition II:

n (B1 B2 ) = 0

For the boundary condition in Eq.(41) k E0 = i j k k sin 0 k cos Ep cos Es Ep sin

= k [iEs cos + j(Ep cos2 + Ep sin2 ) + kEs sin ] B = k E0 = kEs cos kEp kEs sin i j k k sin 0 k cos Rp cos Rs Rp sin

(69) (70)

k R0 =

= k [iRs cos + j(Rp cos2 + Rp sin2 ) + kRs sin ] B = k R0 = kRs cos kRp kRs sin

(71) (72)

k T0 =

i j k k sin 0 k cos Tp cos Ts Tp sin

= k [iTs cos + j(Tp cos2 + Tp sin2 ) + kTs sin ] B = k T0

k Ts cos k Tp k Ts sin

(73) (74)

= B becomes The boundary condition Bz + Bz z

k sin k sin (Es + Rs )ei(k sin xt) = Ts ei(k sin xt) n 1 (Es + Rs ) sin = n 2 Ts sin Es + Rs = Ts We used Snells law in the last equation. This equation is the same as Eq.(68).

(75) (76) (77)

3.3

Condition III:

n (D1 D2 )
1 0 (Ez + Ez ) = 2 0 Ez + 12

For the boundary condition in Eq.(40) (78)

i(k sin xt) i(k sin xt)

1 0 (Ep sin Rp sin )e

= 2 0 (Tp sin )e

+ 12 (r, t)

12 (r, t) = (z )

dk

d12 (k, )ei(kxt) (79)

In Eq.(79) we used the Fourier transformed charge density 12 (k, ). If we apply (1/2 )2 dxdtei(k x t) , then we use (1/4 2 ) dxdtei(k sin xk x) ei( )t = (k sin 9

k ) ( ) and the above equations are held at z = 0

1 0 (Ep sin Rp sin ) (k sin k ) ( ) = 2 0 Tp sin (k sin k )

( ) +

dk d 12 (k , ) (k k ) ( ) (80)

n 1 2 (Ep + Rp ) sin = n 2 2 Tp sin +

12 (k sin , ) 0 From Snell s law 12 (k sin , ) n 1 (Ep + Rp ) = n 2 Tp + 0 n 1 sin

(81) (82) (83)

Here we can neglect the 12 (k sin , ) term 7 .

3.4

Condition IV: t [H1 H2 ] = Js

Hx + Hx = Hx + (Js )x

(84) (85) (86) (87)

Hy +

Hy

Hy

+ (Js )y dkd [Js (k, )]x ei(kxt) , ..

Bx = 0 Hx , ...... [Js (r, t)]x = kEs cos kRs cos i(k sin xt) + e 1 0 1 0 n 1 cos (Es Rs ) 1 kRp kEp + ei(k sin xt) 1 0 1 0 n 1 (Ep + Rp ) 1 = = = =

k Ts cos i(k sin t) e + [Js (r, t)]x 2 0 n 2 cos Ts + c0 [Js (k sin , )]x (88) 2 k Tp i(k sin xt) e + [Js (r, t)]y (89) 2 0 n 2 Tp + c0 [Js (k sin , )]y (90) 2

3.5

Reection and Transmission Coecients

(Ep Rp ) cos = Tp cos 12 (k sin , ) n 1 (Ep + Rp ) = n 2 Tp + 0 n 1 sin n 1 n 2 (Ep + Rp ) = Tp + c0 [Js (k sin , )]y 1 2

In the end we can obtain the following equation for p-wave (91) (92) (93)

and for s-wave Es + Rs = Ts n 1 cos n 2 cos (Es Rs ) = Ts + c0 [Js (k sin , )]x 1 2 (94) (95)

7 2 Laser light(10 mW He-Ne, focused to 20 m, 1026 photons/s/m2 = I/( h ) I = c0 E0 /2 (J/m2 /s = W/m2 ) c=299792458 m, E0 = 1.5 105 V/m ) Bright sunlight (average 480nm, 1018 photons/s/m2 2 E0 = 18V/m.) For electrochemical systems 12 (k sin = 0, = 0) is the order of 10 C/cm = 0.1 2 C/m (fully dissociated 3-mercaptopropionic acid in the 3 3 structure on Au(111) surface = 0.74 C/m2 ). 12 /0 = (1 8) 1010 V/m. However, the surface charge at the electrode in the = 0 limit do not couple to the photon eld at = photon .

10

3.5.1

Usual Solution: No Absorption in the Media 1 and 2.

Here we assume that 12 (k sin , ) = 0, Im( n1,2 ) = 0 (later we will consider the case that the optical constant is complex, i.e. the medium absorb the light.), 1 = 2 = 1, [Js (k sin , )]x = 0, [Js (k sin , )]y = 0. Then for p-wave we can get (Ep Rp ) cos = Tp cos n1 (Ep + Rp ) = n2 Tp and for s-wave Es + Rs = Ts n1 cos (Es + Rs ) = n2 cos Ts

(96) (97)

(98) (99)

If we dene the amplitude reection coecient r and the amplitude transmission coecient for s- and p-waves, rp tp rs ts Rp Ep Tp Ep Rs Es Ts Es (100) (101) (102) (103)

(1 rp ) cos = tp cos n1 (1 + rp ) = n2 tp

(104) (105) (106)

and for s-wave 1 + rs = t s n1 cos (1 rs ) = n2 cos ts We nally get the Fresnel() equations rp = tp = rs = ts = n1 cos + n2 cos n1 cos + n2 cos 2n1 cos n1 cos + n2 cos n1 cos n2 cos n1 cos + n2 cos 2n1 cos n1 cos + n2 cos (109) (110) (111) (112)

(107) (108)

The reectance R is dened as the ratio of the reected power (or ux) to the incident power of the light R= I A cos I n1 R2 /(2c0 ) = = = r2 IA cos I n1 E 2 /(2c0 ) 11 (113)

sin( /2 )
A

= =

A
s

cos A S S cos

Figure 4: Angle dependency of the intensity of light The radiant ux density I (W/m2 ) is given by the averaged Poynting vector < E H >= n E 2 /(2c0 ). In the same way the transmittance T may be given by T = I A cos vt t T 2 cos n2 cos 2 = = t IA cos vi i E 2 cos n1 cos (114)

In Fig.4 and Fig.5 the r, t, R, and T are plotted for the air(1)|water(2) [airwater] and water(1)|air(2) [waterair] interface, respectively. 8 In the both cases ts and tp are always positive, and this means the phase shifts of transmitted wave are always zero. For the case of air(1)|water(2) [airwater] rs is always negative and rp is positive for < p,airwater and become negative for > p,airwater . These means that the phase shift is for reected s wave and phase shift become 0 to at p,airwater . For the case of water(1)|air(2) [waterair] rs is always positive and rp is negative for < p,waterair and become positive for > p,waterair . These means that the phase shift is 0 for reected s wave and phase shift become to 0 at p,waterair . In Appnedix the plot of the phase shift of the reected s and p waves are shown. We also shown the condition of the black lm formation in Appendix. 3.5.2 Brewster angle
9

In Figs.4 and 5, we can nd the point rp = Rp = 0

and this condition is given by sin cos = 0 sin

n1 cos + n2 cos = n1 cos

(115)

(cos sin cos sin ) =

ei + ei ei ei ei + ei ei ei 2 2i 2 2i

8 From Fig.5 (right), intensity of the s-component of the reected wave from the water surface is greater than the p-wave for all the reected angles. Then the polarized sunglass should has better performance if the s-wave component is not transmitted. The s-wave has the electric eld vector in the horizontal direction. Then, the polarized sunglass passes the light with the electric eld vector with the vertical direction. If you try to see the light reected from the sea surface through the polarized sun glass, the intensity of light is lowest and if you rotate the sunglass (or head if you like) 90 degree the reected light becomes the strongest. And when you see the LCD(liquid crystal display) and the watch with liquid crystal through the polarized sunglass, the display becomes totally black if you rotate the glass to 45 degree in the clockwise direction. I think the displayed is designed we can see it even if you wear the polarized sunglass. For the reection from water to air interface as shown in Fig.6, total reection of light R = 1, T = 0 is observed from the angle 48.66 degree. 9 In the laser resonator silica plate is set at the Brewster angle, and the p-wave is transmitted but s-wave is reected and decayed in the resonator.

12

Air(n1=1)|Water(n2=1.332) Interface :He-Ne Laser 1 0.8 0.6 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 -1 0 10 20 30 40 50 60 incident angle (degree) 70 80 90 rs ts tp R or T 1 0.8 0.6 0.4 0.2 0 0

Air(n1 =1)|Water(n2 =1.332) Interface: He-Ne Laser Ts Tp

rp

Rs 10 20 30 40 50 60 70 incident angle (degree)

Rp 80 90

Figure 5: Reection and transmission of He-Ne laser light at air|water interface

Water(n1=1.332)|Air(n2=1) Interface :He-Ne Laser 3 2.5 2 ts R or T 1.5 1 0.5 0 -0.5 0 5 10 15 20 25 30 35 incident angle (degree) 40 45 50 rs rp tp 0.6 0.4 0.2 Rs 0 0 5 10 15 20 25 30 35 incident angle (degree) 40 45 50 Rp 1 Ts 0.8 Tp Water(n1 =1.332)|Air(n2 =1) Interface: He-Ne Laser

Figure 6: Reection and transmission of He-Ne laser light at water|air interface = = Then e2i e2i e2i e2i 4i 4i 2i sin(2 ) 2i sin(2) =0 4i = (116)

2 The angle between the reected light and the transmitted light is orthogonal. n2 sin = n2 sin(/2 ) = n2 cos = n1 sin Brewster = tan1

(117)

n2 (118) n1 From air to water reection, the Brewster angle is 53.10 degree, and from water to air the angle is 36.90 degree.

3.6

Total Internal Reection

Now we will consider the case that the angle between the transmitted wave and the = (n /n ) sin = 1. Here surface normal is 90 degree, i.e., sin c 1 2 c n2 c = sin1 (119) n1 13

and should n2 < n1 . For water to air reection c = 48.66 degree. In this situation the transmitted light is along the surface parallel direction. What happens if > c ? cos = 1 sin2 = i sin2 1 = i (n1 /n2 )2 sin2 1 (120)

The transmitted wave can be written as E = T0 ei(k = T0 e = T0 e = T0 e

r t)

(121) e
k

i(k sin x+k cos z t) i(k sin x t)

(122)
)2 sin2 1z

(n1 /n2

n sin i 2 c x

ei t e
2n2

n2

(123) (124)

(n1 /n2 )2 sin2 1 z c

z c The transmitted light intensity decays as e . The decay length is the order of wavelength because c/ = /2 . The Poynting vector S in the z direction in the medium 2,

(n1 /n2 )2 sin2 1

z S = = = =

1 Re(E B ) z 20 1 Re[E (k E ) ] z 20 1 Re[k |E |2 E (k E ) z 20

=0

(125) (126) (127) (128) (129)

1 Re( 20

z k
k cos =pure imaginary

|E |2 )

= 0

direction in the medium 2 is zero. We call this exponentialThereby the energy ux to z decay wave as evanescent wave, and is used for some interface spectroscopies to detect species located at the evanescent eld. 3.6.1 Phase Shift of the Total Internal Reection Wave

For the total internal reection the amplitudes of the incident wave and the reected wave are the same but there is a phase shift between them. rp = a2 b2 2iab n2 cos n1 cos a ib = = = eip n2 cos + n1 cos a + ib a2 + b2 (130) (131) (132) (133) (134) (135)

a = n2 cos , b = n1 (n1 /n2 )2 sin2 1 2ab tan(p ) = 2 a b2 n1 cos n2 cos c id c2 d2 2icd rs = = = = eis n1 cos + n2 cos c + id c2 + d2 c = n1 cos , 2cd tan(s ) = 2 c d2 d = n2 (n1 /n2 )2 sin2 1

14

1 0.8 phase shift / pi 0.6 rs 0.4 0.2 0 45 50 55 60 65 70 75 80 incident angle (degree) 85 90

rp

Figure 7: Phase shift of the TIR wave (He-Ne laser) from water|air interface.

3.7

Metal Surface

For the metal surface the optical index becomes complex because some part of the light is absorbed by the electronic transition of metallic electrons at Fermi level. n 2 = n2 + i2 , k = n 2 /c = (n2 + i2 )/c From the condition I, n1 sin = n 2 sin = (n2 + i2 ) sin Now is complex, and we dene[13] n 2 cos u2 + iv2 , (u2 + iv2 )2 Here u2 and v2 are real. n2 1 2 2 2 2 = n 2 2 2 2 cos = n 2 (1 2 sin ) = n 2 n1 sin n 2 (138) (139) (137) (136)

For the real and imaginary parts of Eq.(139)

2 2 2 2 2 u2 2 v2 = n2 2 n1 sin

(140) (141)

2u2 v2 = 2n2 2 Then we can get u2 2 =

2 v2 = 2 2 2 n2 2 2 n1 sin + 2 2 2 2 2 2 (n2 2 2 n1 sin ) + 4n2 k2

2
2 2 2 (n2 2 2 n1 sin ) + 2 2 2 2 2 2 (n2 2 2 n1 sin ) + 4n2 k2

(142) (143)

For p wave, rp p eip = = 2 p = n 2 cos n1 cos n 2 2 cos 2 cos n1 n = n 2 cos + n1 cos n 2 2 cos 2 cos + n1 n 2 (n2 2 2 + 2in2 2 ) cos n1 (u2 + iv2 ) 2 (n2 2 2 + 2in2 2 ) cos + n1 (u2 + iv2 )
2 (n2 2 2 ) cos n1 u2 + i(2n2 2 cos n1 v2 ) 2 (n2 2 2 ) cos + n1 u2 + i(2n2 2 cos + n1 v2 ) 2

(144) (145) (146)

15

2 2 2 [(n2 2 2 ) cos n1 u2 ] + (2n2 2 cos n1 v2 ) 2 2 2 [(n2 2 ) cos + n1 u2 ] + (2n2 2 cos + n1 v2 )2

(147)

2 2 2 tan p = {[(n2 2 2 ) cos + n1 u2 ](2n2 2 cos n1 v2 ) [(n2 2 ) cos n1 u2 ] 2 2 2 2 2 2 2 2 2 (2n2 2 cos + n1 v2 )}/[(n2 2 2 ) cos n1 u2 + 4n2 2 cos n1 v2 ] 2 2n2 u2 2 (n2 2 k2 )v2 2n2 cos 2 (148) 2 2 2 2 2 (n2 + 2 2 ) cos n1 (u2 + v2 ) 2n1 cos 2n1 n 2 cos p eip = = (149) n1 cos + n 2 cos n1 n 2 cos + n 2 2 cos 2n1 n2 cos + i2n1 2 cos (150) 2 n1 u2 + (n2 2 2 ) cos + i(n1 v2 + 2n2 2 cos ) 2 n2 2 + 2 2 4n2 (151) 1 cos = 2 2 2 [n1 u2 + (n2 2 2 ) cos ] + (n1 v2 + 2n2 2 cos ) 2 2 n1 2 u2 + 2 (n2 2 2 ) cos n1 n2 v2 2n2 2 cos 2n1 cos (152) 2 2 2 2n2 cos [n2 1 u2 + n1 (n2 2 ) cos + n1 2 v2 + 2n2 2 cos ]

= tp = =
2 p =

tan p =

The tp in the book Born and Wolf, Principles of Optics (7th ed) p.575 Eqs.(14) and (15)[13] may be wrong. For s wave rs = s eis = 2 s = tan s = ts =
2 s =

tan s =

n1 cos n 2 cos n1 cos (u2 + iv2 ) = n1 cos + n 2 cos n1 cos + u2 + iv2 2 2 n1 cos u2 iv2 (n1 cos u2 )2 + v2 = 2 n1 cos + u2 + iv2 (n1 cos + u2 )2 + v2 2v2 n1 cos 2 2 n1 cos2 u2 2 v2 2n1 cos 2n1 cos s eis = = n1 cos + n 2 cos n1 cos + u2 + iv2 2 2 4n2 2n1 cos 1 cos = 2 n1 cos + u2 + iv2 (n1 cos + u2 )2 + v2 2v2 n1 cos v2 2 = 2 n1 cos + u2 2n1 cos + 2u2 n1 cos

(153) (154) (155) (156) (157) (158)

For He-Ne laser (632.8 nm) the reectance at the air|gold surface (n = 0.181, = 2.99) is shown in Fig.7. For IR light at 3100 nm the reectance at the air|gold surface (n = 1.728, = 19.2) is shown in Fig. 8. The intensities of light in the z - and y -component per unit area on the Au surface becomes
2 2 (Ep /Ep )2 / cos |E0z + R0z |2 /Ep / cos = | Ep sin Rp sin |2 /Ep / cos 2 = | Ep sin rp Ep sin |2 /Ep / cos 2 = | Ep sin (p cos p + ip sin p )Ep sin p |2 /Ep / cos

= sin 2 (1 + 2p cos p + 2 p )/ cos

2 (Es /Es )2 / cos |E0y + R0y |2 /Es / cos 2 2 = |Es + Rs |2 /Es / cos = |Es + rs Es |2 /Es cos

(159)

= (1 + 2s cos s + 2 s )/ cos

(160)

These equations give the basics of the surface sensitivity of the RAIRS(Reection Absorption Infrared Spectroscopy) and Polarization-Modulation FTIR spectroscopy, and the 16

1 0.98 0.96 reflectance 0.94 0.92 0.9 0.88 0.86 0 10 20 30 40 50 60 70 incident angle (degree) 80 90 Rp R
s

250 s-wave 200 phase shift (degree) 150 100 50 0 0 10 20 30 40 50 60 70 incident angle (degree) 80 90 p-wave

Figure 8: He-Ne laser (632.8 nm) reection from air|Au surface

1 0.98 0.96 reflectance 0.94 0.92 0.9 0.88 0.86 0.84 0.82 0 10 20 30 40 50 60 70 incident angle (degree) 80 90 Rp Rs phase shift (degree) 200 180 160 140 120 100 80 60 40 20 0 0 10 20 30 40 50 60 70 incident angle (degree) 80 90 p-wave s-wave

Figure 9: IR light (3100 nm) reection from air|Au surface angle dependence are shown in Fig.9. The 180 degree phase change of s-wave leads to destructive interference and no interaction with surface dynamic dipoles from molecular vibrations. 10 For the surface normal component of the p-wave the interference is constructive and it can excite the dynamic dipole perpendicular to the surface. The excitation is ecient for a higher angle of incidence.

Surface Plasmon

The electronic charges on metal boundary can perform coherent uctuations which are called surface plasma oscillations. The uctuations are conned at the boundary and vanishes both sides of the metal surface. This plasmon waves have p-character because the surface charge induce the discontinuity of the electric eld in the surface normal z -direction, but s-waves has only Ey component (no Ez component). Now we consider the air(medium 2)|metal(medium 1) surface where the electric elds are dumped both side of the interface. Using a pure imaginary kz 2 the electric and magnetic eld in medium 2(air, z > 0)

10

The surface parallel x-component (not shown here) of p-wave is also negligible.

17

45 40 35 30 25 20 15 10 5 0 0 10 20 30 40 50 60 incident angle (degree) 70 80 90 s-parallel Eq.(160) p-normal Eq.(159)

/E )2 / cos and (E /E )2 / cos . The IR light is 3100 nm and Figure 10: plots of (Ep s p s

the IR light is reected from the air|Au surface. (Es /Es )2 / cos is negligible because the phase shift is almost 180 degree between the incident wave and the reected wave. can be given by E2 Ex2 = 0 ei(kx2 x+kz2 z t) Ez 2 0 = Hy2 ei(kx2 x+kz2 z t) 0

(161)

H2

(162)

Using a pure imaginary kz 1 the electric and magnetic eld in medium 1(metal, z > 0) can be given by E1 Ex1 = 0 ei(kx1 xkz1 z t) Ez 1 0 = Hy1 ei(kx1 xkz1 z t) 0

(163)

H1

(164)

From the Condition I, we can get kx1 = kx2 = kx Ex1 = Ex2 Ez 1 = Ez 2 From condition IV, Hy 1 = Hy 2 here we assume (Js )y 0 From condition III, 1 0 Ez 1 = 2 0 Ez 2 here we assume 18 12 (kx , ) D1z , D2z (170) (171) (168) (169) (165) (166) (167)

From Eq.4 and J 0 rotH = rotH = D t i

x

(172) j
y

k
z

= i

Hyi

Hyi + k Hyi z x

(173)

From the i component of the above equation, ikz 1 Hy1 = i0 1 Ex1 ikz 2 Hy2 = i0 2 Ex2 Ex1 = Ex2 then kz 1 kz 2 Hy 1 + Hy2 = 0 0 1 0 2 kz 1 kz 2 + =0 1 2 From Eqs.(21) and (26)
2 2 k 2 = kx + kzi = i 2 kx = 1 2 kx = 2

(174) (175)

(176)

Hy1 = Hy2 then (177)

(178) (179)

= 2

c c c

2 kz 1 2 kz 2

2 kz 1 1
2

(180)

From the last two equations Eqs.(179) and (180).

2 kx =

1 2 1 + 2 2 1 1 + 2 2 2 1 + 2

c c c

(181)
2

2 kz 1 = 2 kz 2 =

(182)
2

(183)

If we remind that 1 = 2 = 2 1 + i1 , 2 kx =

c c

( 1 + i1 )2 ( 1 + i1 ) + 2

=
If we assume 1 < |1 |

2 1 (1 + 2 ) + 1 + i[1 ( 1 + 2 ) 1 1 ] 2 2 ( 1 + 2 ) + 1

(184)

Re(kx ) = Im(kx ) =

c c

1 2 1 + 2 1 2 1 + 2 19

1/2

(185)
3/2

1 212

(186)

The surface plasmon decay in x-direction can be evaluated from Im(kx ) because the intensity decreased as exp[2Im(kx )x]. The decay length L12 may be obtained as L12 = [2Im(kx )]1 = c 1 2 1 + 2
3/2

12 1

(187)

For the water|metal interface the decay lengths L12 are 6.4 m for gold (16.6 m for air|gold surface), 12.3 m for silver, and 5.5 m for aluminum. The decay length L12 is the key parameter to carry out a SPR imaging measurements 11 . In addition there is a temporal decay in , please refer the Raethers book for details[7]. The dispersion relation kx vs become close to the light line 2 /c at small kx , because in the limit that 0, 1 >> 2 . At large kx the denominator of Eq.(185) becomes zero (188) 1 + 2 = 0 For simple metals the dielectric constant is given by the plasma frequency p [16] 1 =1
12 2 p (189) 2 From Eqs.(188) and (189) the surface plasma frequency sp may be obtained as

sp = p

1 1 + 2

(190)

In Fig.10 we plot the dispersion relation Eq.(185).

3.5 bulk plasmon energy 3 2.5 energy (eV) 2 1.5 1 0.5 0 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 kx (108 m-1) surface plasmon energy

Figure 11: Surface plasmon dispersion (kx ) on gold surface. The vertical axis is scaled as h (eV). The straight solid line in the gure shows the light line kx = 2 /c. The energy of bulk plasmon is 3.22 eV, and that of surface plasmon is 2.28 eV and shown as the arrows in the gure. In the z -direction the electric eld of the surface plasmon decays as Ez e|kzi ||z | . If we assume 1 < |1 | again,
2 kz 1

12 1 + 2 1 2 2 + 2

c c

(191)
2

2 kz 2
11

(192)

From this discussion the lateral resolution of SPR is more than micrometer size. Aluminium has small L12 and living cell is observed with SPR imaging. [15] 12 If the energy loss of electron beam reected from the metal surface is measured, the plasma frequency can be determined. Surface plasma frequency can be also measured in the same way.

20

 1 + 2 < 0 then kzi is purely imaginary. For He-Ne laser light (632.8 nm) on the gold surface 1/Re(kz 1 )(metal) = 32 nm, 1/Re(kz 2 )(air) = 285 nm (193) From the k component of rotH = D/t [cf. Eq.(173)] Hy1 x Hy2 x = ikx1 Hy1 = i0 1 Ez 1 = ikx2 Hy2 = i0 2 Ez 2 (194) (195)

From the i component of Eqs.(172-173) and Eq.(194) Hy 1 = Ez 1 Ex1 Hy 2 Ez 2 Ex2 = 0 0 1 Ex1 1 Ez 1 = kz 1 kx1 (196) (197) (198) (199) (200)

kx1 kz 1 0 0 2 Ex2 2 Ez 2 = = kz 2 kx2 kx2 = kz 2

5
5.1

Excitation of Surface Plasmon by Light

ATR (Attenuated Total Reection) Coupler Method

We will consider the situation that the light is reected from a metal surface covered with a dielectric medium (pr > 1) , e.g. with a BK7 half cylinder glass prism (n=1.515 at 633 nm) or SF10 glass prism (n=1.723 at 643.8 nm). The x and z components of the wavevector in the prism are given by pr (201) kx = pr sin pr = npr sin pr c c pr (202) kz = pr cos pr = npr cos pr c c Here pr means the prism. The resonance condition of the light in the prism with the surface plasmon at metal(1)|air(2) interface (Kretschmann-Raether conguration) is 13
pr sp kx = kx

(203) c (204)

pr

sin pr = c

1 2 1 + 2

sp . Here we use Eq.(181) for kx

13 In the ATR coupler there are two types, one is prism|metal lm|sample Kretschmann-Raether type and the other is prism|thin sample(thickness is wavelength of light)|metal Otto type. In the Otto type the controll of the thin gap is dicult, and then Kretschmann-Raether type is usually used. However, the Au lm thickness should be controlled around 50 nm in the Kretschmann-Raether type, otherwise the SPR resonance (T 0) become un-sharp. In the electrochemical environment the gold lm is not stably deposited on the glass plate when the potential is applied. Then (3-mercaptopropyl)trimethoxysilane(MPS) is used for the binder between the glass and gold (see Appendix). Otto rst showed that light can excite surface plasmon by the use of ATR coupler[3], but He claimed that the surface plasmon can not be excited by the Kretschmann-Raether type ATR coupler and later it was found that his claim was wrong. After that Otto never went mainstream in the SPR world!?

21

pr kxpr d1 pr 2
y z x

Figure 12: Schematic diagram of ATR coupler: Kretschmann-Raether type 5.1.1 Prisim|Metal|Medium Three-Layer Model

The reectivity Rpr|1|2 may be given by Frensels equations of the prism|metal|air threelayer system.
zi zk n k k i k n k cos i n i cos k ki n kk = = kzk zi n k cos i + n i cos k n k k + n i ki kk

p rik =

n k kzi c n i n k kzi c n i

n i kzk c n k n i kzk c n k

(205)

kzi n 2 i kzi n 2 i

kzk n 2 k kzk n 2 k

kzi / i kzk / k kzi / i + kzk / k

(206) (207)

p p rki = rik

For transmission tp ik = tp ki
p tp ik tki

n i p (1 + rik ) n k n k n k p p = (1 + rki )= (1 rik ) n i n i p p = (1 + rik )(1 rik )

(208) (209) (210)

The total reection of the three-layer model becomes R =

p rik = p |rpr12 |2 p p 2ikz1 d1 rpr1 + r12 e = p p 2ikz1 d1 1 + rpr1 r12 e 2

(211) (212)

npr sin pr

cos i /n i cos k /n k n k cos i n i cos k = n k cos i + n i cos k cos i /n i + cos k /n k = n 1 sin 1 = n 2 sin 2
2 2 1/2 / 1/2 / k i ( k n 2 ( i n2 pr sin pr ) pr sin pr ) 2 2 1 / 2 1 / 2 2 2 ( i npr sin pr ) / i + ( k npr sin pr ) / k 2 1/2 / 1 cos pr /npr ( 1 n2 pr sin pr ) 2 1/2 / cos pr /npr + ( 1 n2 1 pr sin pr )

2 1/2 2 2 1/2 n k cos k = n k (1 sin2 k )1/2 = n k (1 n2 2 = ( n2 pr sin pr /n k) k npr sin pr ) p rik = p rpr1

(213) (214)

22

The above equation for R can be understood by considering

p rpr12

p p 2ikz1 d1 rpr1 + r12 e p p 2ikz1 d1 1 + rpr1 r12 e

(215)

p p 2ikz1 d1 p p 2ikz1 d1 p p 2 4ikz1 d1 (rpr1 + r12 e )(1 rpr1 r12 e + (rpr1 )2 (r12 ) e ....) p p 2ikz1 d1 p p 2ikz1 d1 = rpr1 + r12 e (rpr1 )2 r12 e p p 2 4ikz1 d1 p p 2 4ikz1 d1 rpr1 (r12 ) e + (rpr1 )3 (r12 ) e + ... p p p 2ikz1 d1 p p p 2 4ikz1 d1 = rpr1 + (1 rpr1 )2 r12 e + rpr1 [(rpr1 )2 1](r12 ) e + ... p p p p = rpr1 + (1 + rpr1 )r12 (1 rpr1 ) e2ikz1 d1
p p tp pr1 r12 t1pr

p p p p p + (1 + rpr1 )r12 (rpr1 )r12 (1 rpr1 ) e4ikz1 d1

p p p p tp pr1 r12 r1pr r12 t1pr

p p p p p p p 2ikz1 d1 4ikz1 d1 = rpr1 + tp + tp pr1 r12 t1pr e pr1 r12 r1pr r12 t1pr e

(216) (217)

phase factor

kz 1 d1 = k1 (d1 cos 1 ) is optical path length.

1/2

n2 pr kz 1 d1 = k1 d1 cos 1 = n 1 d1 1 2 sin2 pr c n 1 2 1/2 d1 ( 1 n 2 = pr sin pr ) c

1 0.9 0.8 0.7 reflectance 0.6 0.5 0.4 0.3 0.2 0.1 0 35 36 37 38 39 40 41 42 incident angle (degree)

(218)

43

44

45

Figure 13: SPR curve for SF10(n = 1.723)|gold(50nm, n 1 = 0.1726+i3.4218) |air(n = 1.0) for He-Ne laser light (633 nm). The FORTRAN program to get the SPR curve by Eq.(211) is c234567-- spr_angle_3layer.f --c complex calculation implicit real*8 (a-h,o-z) complex*16 e1,rpr1,r12,rpr12,rpr12c,alpha,ref complex*16 aaa,fukso,e2

23

c c c

c c

c=2.99792458d8 hbar=6.5822d-16 pi=acos(-1.0d0) ------ air ---e2=dcmplx(1.0d0,0.0d0) ----- SF10 633 nm enpr=1.723d0 ----- gold 633 nm e1n=0.1726d0 e1k=3.4218d0 e1r=e1n**2-e1k**2 e1i=2.0d0*e1n*e1k e1=dcmplx(e1r,e1i) ---- gold thickness (m) d1=50.0d-9 ramd=633.0d-9 omega=2.0d0*pi/ramd*c fukso=dcmplx(0.0d0,1.0d0) write (6,*) ramd,omega,hbar*omega --------- angle scan ---------ang0=35.0d0 ang1=45.0d0 do i=1, 1001 theta=(ang0+dble(i-1)/1000.0d0*(ang1-ang0))/180.0d0*pi rpr1=(cos(theta)/enpr & sqrt(e1-enpr**2*sin(theta)**2)/e1) & / (cos(theta)/enpr + & sqrt(e1-enpr**2*sin(theta)**2)/e1) r12=( sqrt(e1-enpr**2*sin(theta)**2)/e1 sqrt(e2-enpr**2*sin(theta)**2)/e2 ) / ( sqrt(e1-enpr**2*sin(theta)**2)/e1 + sqrt(e2-enpr**2*sin(theta)**2)/e2 ) aaa=2.0d0*omega/c*d1*sqrt(e1-enpr**2*sin(theta)**2) alpha=aaa*fukso rpr12=(rpr1+ r12*exp(alpha))/(1.0d0+rpr1*r12*exp(alpha)) rpr12c=conjg(rpr12) ref=rpr12*rpr12c write (6,*) theta/pi*180.0d0,dble(ref) enddo end

& & &

The calculated results are shown in Fig. 12. At resonance or R = 0 the power of the SPs is lost by internal absorption in the metal. This loss is compensated by the power of the incoming light. Both have to be equal in the steady state. If the reectivity R has lowest value, the intensity of the electromagnetic eld reaches its maximum at the metal surface. For 600 nm light the maximum enhancement of the electric eld intensity is ca. 200 for silver lm (60 nm thickness), 30 for gold lm, 40 for aluminum lm, and 7 for copper lm, respectively[7].

24

5.2

General solution of N-layer model. [ F. Abel` es, Ann. Phys. (Paris) 5, (1950) 596. W. N. Hansen, J. Opt. Soc. Amer. 58 (1968) 380. ]

z1 1 2

z2

z3

zN-2 zN-1

3 4 N-2 N-1 N

Figure 14: N-layer model for SPR measurement. The tangential elds at the rst boundary z = z1 = 0 are related to those at the nal boundary z = zN 1 by U1 V1 = M2 M3 ...MN 1 UN 1 VN 1 =M UN 1 VN 1 (219)

For p-wave at boundary k ,

T R Uk = Hy + Hy

(220) (221)

Vk = and Mk =

T Ex

R Ex

cos k i sin k /qk iqk sin k cos k

(222) (223)

1/2 Here qk = (k / cos k k) k = 1

2 2 1/2 ( k n1 sin 1 ) 1/2 cos k = qk (224) = (1/ k) k 2 2 2 1/2 n k cos k (zk zk1 ) = (zk zk1 ) ( k n 2 (225) k = 1 sin 1 ) 0 0

The reection and transmission coecient for p-wave (TM) is rp = Mij =

k=2

(M11 + M12 qN )q1 (M21 + M22 qN ) (M11 + M12 qN )q1 + (M21 + M22 qN )
N 1

(226) (227) (228) (229) (230) (231) (232) (233) (234)

Mk
ij

i, j = 1, 2

Rp = |rp |2 rp = r p tp H tp E = arg(r ) = =
1/2 ir Rp e p p

Tp = t p =

2q1 (M11 + M12 qN )q1 + (M21 + M22 qN ) N n1 p t 1 n N H N Re( nN cos N /n 2 N) p 2 |tH | 2 1 n1 cos 1 /n1 arg(tp E) 25

For s-wave (TE) the above equations hold except qk = k cos k k (235)

In this sense these equations are easy to calculate the multilayer optical problem.

1 0.9 0.8 0.7 reflectance 0.6 0.5 0.4 0.3 0.2 0.1 0 35 36 37 38 39 40 41 42 incident angle (degree) 43 44 45 (II) SAM on Au (I) Au

Figure 15: SPR curves for (I) the SF10 glass prism(n=1.723)|Au(n + ik = 0.1726+ i3.4218, 50nm)|Air(n=1.0) and (II) the SF10 prism|Au(50nm)|SAM(n=1.61245, 1nm)|Air systems. The SPR resonance calculation FORTRAN program for N-layer system is given in the following, and the results for the 4-layer system [ prism|gold|Self-AssmbledMonolayer(SAM, 1 nm thickness)|Air ] is shown in Fig. 14. The SAM lm thickness of 1 nm is clearly seen as the shift of the SPR angle. c234567-- spr_angle_Nlayer.f --c 1 | 2 | 3 ... N-2|N-1|N :N layer system c complex calculation implicit real*8 (a-h,o-z) parameter (nlay=10) complex*16 e(nlay) , em(nlay,2,2) ,emtot(2,2) complex*16 emtot1(2,2) dimension en(nlay), ek(nlay), d(nlay) complex*16 beta,q,rp,q1,qn,ref,tp,tra complex*16 fukso c=2.99792458d8 hbar=6.5822d-16 26

field enhancement

25 20 15 10 5

35

36

37

incident angle (degree)

for

38

39

40

41

42

43

44

45

2 Figure 16: Field enhancement factor |tp h| prism|Au(50nm)|SAM(n=1.61245, 1nm)|Air system.

the

the

SF10

c c

c c

pi=acos(-1.0d0) ------- N layer ---nlayer=4 ------ air ---en(4)=1.d0 ek(4)=0.0d0 er=en(4)**2-ek(4)**2 ei=2.0d0*en(4)*ek(4) e(4)=dcmplx(er,ei) ----- SF10 633 nm en(1)=1.723d0 ek(1)=0.0d0 er=en(1)**2-ek(1)**2 ei=2.0d0*en(1)*ek(1) e(1)=dcmplx(er,ei) ----- gold 633 nm en(2)=0.1726d0 ek(2)=3.4218d0 er=en(2)**2-ek(2)**2 ei=2.0d0*en(2)*ek(2) e(2)=dcmplx(er,ei) ---- gold thickness (m) d(2)=50.0d-9 ------ SAM ----------en(3)=1.61245 ek(3)=0.0d0 er=en(3)**2-ek(3)**2 ei=2.0d0*en(3)*ek(3) e(3)=dcmplx(er,ei) ----- SAM thickness ---

27

d(3)=1.0d-9 c ramd=633.0d-9 omega=2.0d0*pi/ramd*c fukso=dcmplx(0.0d0,1.0d0) write (6,*) ramd,omega,hbar*omega --------- angle scan ---------ang0=35.0d0 ang1=45.0d0 do i=1, 1001 theta=(ang0+dble(i-1)/1000.0d0*(ang1-ang0))/180.0d0*pi q1=sqrt(e(1)-en(1)**2*sin(theta)**2)/e(1) qn=sqrt(e(nlayer)-en(1)**2*sin(theta)**2)/e(nlayer) do j=2, nlayer-1 beta=d(j)*2.0d0*pi/ramd*sqrt(e(j)-en(1)**2*sin(theta)**2) q=sqrt(e(j)-en(1)**2*sin(theta)**2)/e(j) em(j,1,1)=cos(beta) em(j,1,2)=-fukso*sin(beta)/q em(j,2,1)=-fukso*sin(beta)*q em(j,2,2)=cos(beta) enddo emtot(1,1)=dcmplx(1.0d0,0.0d0) emtot(2,2)=dcmplx(1.0d0,0.0d0) emtot(1,2)=dcmplx(0.0d0,0.0d0) emtot(2,1)=dcmplx(0.0d0,0.0d0) do j=2, nlayer-1 emtot1(1,1)=em(j,1,1) emtot1(1,2)=em(j,1,2) emtot1(2,1)=em(j,2,1) emtot1(2,2)=em(j,2,2) emtot=matmul(emtot,emtot1) enddo rp=( (emtot(1,1)+emtot(1,2)*qn)*q1 (emtot(2,1)+emtot(2,2)*qn) ) / ( (emtot(1,1)+emtot(1,2)*qn)*q1 + (emtot(2,1)+emtot(2,2)*qn) ) tp=2.0d0*q1/( (emtot(1,1)+emtot(1,2)*qn)*q1 + & (emtot(2,1)+emtot(2,2)*qn) ) & & & rp=( (emtot(1,1)+emtot(1,2)*qn)*q1 (emtot(2,1)+emtot(2,2)*qn) ) / ( (emtot(1,1)+emtot(1,2)*qn)*q1 + (emtot(2,1)+emtot(2,2)*qn) ) tp=2.0d0*q1/( (emtot(1,1)+emtot(1,2)*qn)*q1 + (emtot(2,1)+emtot(2,2)*qn) ) ref=rp*conjg(rp) enh=tp*conjg(tp) tra=tp*conjg(tp)/cos(theta)*en(1)*dble(qn) 28

& & & &

write (6,*) theta/pi*180.0d0,dble(ref),enh write (6,*) theta/pi*180.0d0,dble(ref),dble(tra) enddo end

Acknowledgement

The author would like to thank for Mr. A. Shirakami for his great contribution to the startup of the SPR study in our laboratory. The author also thank Prof. R. Corn, Prof. T. Kakiuchi and Dr. D. Hobora for helpful discussions on this subject.

Appendix

In Fig.16 the phase shifts of the reected wave(He-Ne laser light) are shown in the case of Air(n = 1) Water(n = 1.332)(top gure) and WaterAir(bottom gure). For the case of WaterAir the critical angle for total reection is 48.66 degree and the phase shift is shown in Fig.6.

References
[1] R. H. Ritchie. Plasma losses by fast electrons in thin lms. Phys. Rev., 106:874881, 1957. [2] R. W. Wood. On a remarkable case of uneven distribution of light in a diraction grating spectrum. Phil. Mag. Ser. 6, 4:396402, 1902. [3] A. Otto. Excitation of surface plasma waves in silver by the method of frustrated total reection. Z. Physik, 216:398410, 1968. [4] H. Raether E. Kretschmann. Radaitive decay of non-radiative surface plasmons excited by light. Z. Naturforsch., 23A:21352136, 1968. [5] J. G. Gordon II and J. D. Swalen. The eect of thin organic lms on the surface plasma resonace on gold. Opt. Commun., 22:374376, 1977. [6] C. Nylander B. Liedberg and I. Lundstrom. Surface plasmon resonance for gas and biosensing. Sensers and Acutuators, 4:299304, 1983. [7] H. Raether. Surface Plasmons on Smooth and Rough Surface and on Gratings. SpringerVerlag, Berlin, 1988. [8] B. L. Frey R. Corn D. G. Hanken, C. E. Jordan. Surface plasmon reasonance measurements of ultrathin organic lms at electrode surfaces. Electroanal. Chem., 20:141225, 1998. [9] W. Knoll. Interfaces and thin lms as seen by bound electromagnetic waves. Annu. Rev. Phys. Chem., 49:569638, 2000. [10] R. Corn J. M. Brockman, B. P. Nelson. Surface plasmon reasonance imaging measurements of ultrathin organic lms. Annu. Rev. Phys. Chem., 51:4163, 2000. [11] D. M. Kolb. The study of solid-liquid interfaces by surface plasmon polariton excitation. In V. M. Agranvich and D. L. Mills, editors, Surface Polaritons, chapter 8, pages 299329. North-Holland, Amsterdam, 1982. [12] V. Scheumann Z. Zizlsperger J. Mack G. Jung A. Badia, S. Arnold and W. Knoll. Probing the elecrochemical deposition and/or desorption of self-assembled and electropolymerizable organic thin lm by surface plasmon spectroscopy and atomic force microscopy. Sensors and Actuators B, 54:145165, 1999. [13] M. Born and E. Wolf. Principles of Optics (7th expanded edition). Cambridge University Press, Cambridge, 1999.

29

[14] E. Hecht. Optics. Addison-Wesley, San Francisco, 2002. [15] S. Herminghaus M. Riedel P. Leiderer U. Weiland K.-F. Giebel, C. Bechinger and M. Bastmeyer. Imaging of cell/substrate contacs of living cells with surface plasmon resonance microscopy. BioPhys. J., 76:509516, 1999. [16] C. Kittel. Introduction to Solid State Physics (7th edition). Wiley, New York, 1996.

30

180

rs
150

phase shift (degree)

120

90

air to water

rp

60

30

0 0 10 20 30 40 50 60 incident angle (degree) 70 80 90

180

150

water to air
phase shift (degree) 120

90

rp

60

30

rs
0 0 10 20 30 incident angle (dgree) 40

Figure 17: Phase shifts of the reected wave(He-Ne laser light) are shown in the case of Air(n = 1) Water(n = 1.332)(top gure) and WaterAir(bottom gure). For the case of WaterAir the critical angle for total reection is 48.66 degree and the phase shift is shown in Fig.6. In Fig.17 schematic diagram of light reection from black lm is shown. When the incident angle is small, the phase shift of s wave is for OC and zero for OAB and the phase shift of p wave is zero for OC and for OAB.(Please see Fig.16) If the lm thickness d is much smaller than the wavelength/4, the interference between OC and wave from B become destructive.(Please note that the formulation is shown in the inset of Fig.17). Then we can see no light from thin lm like lipid bilayer or soap bubble. 31

Figure 18: Schematic diagram of light reection for black lm. When the incident angle is small, the phase shift of s wave is for OC and zero for OAB and the phase shift of p wave is zero for OC and for OAB.(Please see Fig.16) If the lm thickness d is much smaller than the wavelength/4, the interference between OC and wave from B become destructive. Please note that the formulation is shown in the inset. Then we can see no light from thin lm like lipid bilayer or soap bubble.

32

Appendix
Home-made SPR Apparatus (by A. Shirakami)
He-Ne laser 628.3 nm Beam splitter ND filter Rotating stage

Polarizer

Lens Lens Detector

2

PC

Lens Detector

A/D converter

Electrochemical Cell for SPR

SF-10 hemicylindrical prism Working electrode Potentiostat Reference electrode

Glass slide Gold layer

Counter electrode

Details of the Sample Setup

SF10 glass Prism n matching oil SF10 slide glass MPS Au film (50nm) SAM
Figure 19: 33