Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
1039/9781847558220-00015
CHAPTER 2
Resonance
ROB P.H. KOOYMAN
2.1 Introduction
In the last two decades, surface plasmon resonance (SPR) has evolved from a
fairly esoteric physical phenomenon to an optical tool that is widely used in
physical, chemical and biological investigations where the characterization of
an interface is of interest. Recently, the field of SPR nano-optics has been added
where metallic structures on a nanoscale can be designed such that they can
perform certain optical functions. This chapter will be mainly concerned with
the more conventional, well-understood SPR theory used in sensor applications
and it will touch upon some of the newer developments relevant for this area.
Essential for the generation of surface plasmons (SPs) is the presence of free
electrons at the interface of two materials – in practice this almost always
implies that one of these materials is a metal where free conduction electrons
are abundant. This condition follows naturally from the analysis of a metal-
dielectric interface by Maxwell’s equations. From this analysis, the picture
emerges that surface plasmons can be considered as propagating electron density
waves occurring at the interface between metal and dielectric. Alternatively,
surface plasmons can be viewed as electromagnetic waves strongly bound to
this interface; it is found that the surface plasmon field intensity at the interface
can be made very high, which is the main reason why SPR is such a powerful
tool for many types of interface studies.
Experimental research on SPs started with electron beam excitation; in 1968,
optical excitation was demonstrated by Otto [1] and Kretschmann and Raether [2].
This last approach turned out to be much more versatile, so in this chapter the
focus will be on the optics of SPR. The following is by no means intended as an
15
View Online
16 Chapter 2
in-depth treatment of surface plasmons, rather it is an attempt to provide a low-
threshold introduction to the physics of SPR for those who are actually
involved in SPR work and want to understand a bit more than ‘‘measuring
the shift of the SPR dip’’.
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
eq. (2.1), we conclude that for this case in medium 2 there is only a traveling wave
parallel to the interface:
E2 ¼ E0 eky2 y expðjot jkx xÞ ð2:5Þ
with the amplitude of the electric field exponentially decaying along the y-direction
with a characteristic distance 1/ky2 1/jky2. For obvious reasons, this field in
medium 2 is denoted as the evanescent field. Eq. (2.4) can be used to calculate its
penetration depth, which is of the order of half a wavelength. This explains the
interface sensitivity of the evanescent field: only close to the interface is an
electromagnetic field present; therefore, only a changing dielectric property (e.g.
a changing refractive index) in the vicinity of the interface will influence this field.
We will see that also in SPR an evanescent field is generated.
18 Chapter 2
light interacting with an interface. For any interface between two media, the
complex reflection coefficient rp for p-polarized incident light electric field is
described by Fresnel’s equations (see, e.g., ref. [5] for a derivation on the basis
of Maxwell’s equations):
Ei jj tanða bÞ jj
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
rp ¼ ¼ rp e ¼ e ð2:6aÞ
Er tanða þ bÞ
where Ei and Er are the incident and reflected electric fields, respectively, and
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
Now, following Cardona [4], two special cases exist: if a+b ¼ p/2, then the
denominator of eq. (2.6a) becomes very large and thus Rp becomes zero. This
situation describes the Brewster angle, where there is no reflection for
p-polarized light. The other special case occurs when ab ¼ p/2: we see from
eqs. (2.6a) and (2.6b) that Rp becomes infinite: there is a finite Er for a very
small Ei. This circumstance corresponds to resonance. From this relation
between a and b we can deduce the dispersion relation if ab ¼ p/2, then
cosa ¼ sinb and tana ¼ k1x/k1y ¼ n2/n1. For the components of the wave-
vector k ¼ (kx, ky), we can write
e1
k2x ¼ k21 k2y1 ¼ k21 k2x ð2:7Þ
e2
rffiffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
o e1 e2 o e2i
kx ¼ and kyi ¼ ð2:8Þ
c e1 þ e2 c e1 þ e2
o2p
e2 ðoÞ ¼ 1 ð2:9aÞ
o2
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
op ¼ 4pne e2 =me ð2:9bÞ
3
Note that in Figure 2.1 the direction y (instead of z) is perpendicular to the surface.
4
Dielectric constant and refractive index are related by e n2.
View Online
depths for a real case, on the basis of eq. (2.8). We take l ¼ 700 nm, thus
o ¼ 2.69 1015 s1 and a gold/water interface. At this wavelength egoldE 16
and ewaterE1.77. We calculate for the penetration depths 1/ky,water ¼ 238 nm
and 1/ky,gold ¼ 26 nm.
Now all ingredients are available to appreciate the use of SPR in sensor
applications. Let us assume that we have a situation where molecules X are
allowed to adsorb to the water/metal interface. We can view this as a process
where water molecules are replaced by molecules X. Because, generally,
eXaewater, the average dielectric constant close to the interface will change.
Equation (2.8) then describes the concomitant change of the wavevector kx.
Because the SP field is evanescent in the direction perpendicular to the inter-
face, a change of the dielectric constant e2 is only detectable in SP character-
istics if this change occurs within the penetration depth of the SP field: an SPR
sensor will only be sensitive to molecular processes (binding, adsorption, etc.)
that occur at a distance to the metal surface that is roughly half the wavelength
of the used light.
20 Chapter 2
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
Figure 2.2 Dispersion relation for surface plasmons. Curves I and II represent the SP
dispersion for the interfaces e3/em and e1/em, respectively. The lines a and b
are the dispersion relations for ‘‘normal’’ light in medium e3 and e3,
respectively, which are dependent on the angle of incidence a in the
experimental setup as indicated in the inset. By varying a, any line c
between the lines a and b can be realized.
match the wavevector necessary for SP excitation. In this way, any kx between
the two lines, labeled a and b in Figure 2.2, can be set. As an example, one such
line, labeled c, is indicated. This so-called attenuated total reflection (ATR)
technique was first demonstrated5 by Kretschmann and Raether [2] and has
since then almost become the standard technique for SP excitation.
Another way of providing a wavevector appropriate for SP excitation is the
use of a metal layer on which a periodic structure is prepared [6] as illustrated in
Figure 2.3.
When light with wavevector kx ¼ 2p/l.nisiny falls on such a structure, this
acts as a diffraction grating and diffraction orders m ¼ 0, 1, 2, . . . are
generated in the reflected light (see, e.g., ref. [7]). The generated wavevector
kx,net parallel to the interface can be written as
2p
kx;net ¼ kx þ m ð2:10Þ
L
5
In fact, Otto was the very first to demonstrate this in a somewhat less versatile form.
View Online
where L is the periodicity of the grating. Again, the wavevector kx,net can be
tuned to the SPR wavevector, given by eq. (2.8),6 by changing the incident
angle.
Up to now the required polarization direction of the incoming light remained
unmentioned. As already pointed out, SPs are conductivity fluctuations
brought about by collective surface charge density oscillations. These charge
density waves have to be excited by an external electric field. Only an electric
field with a component perpendicular to the interface can induce a surface
charge density; only p-polarized light has a perpendicular electric field
component.
22 Chapter 2
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
Figure 2.4 Field enhancement for various values of the thickness of the gold layer.
Wavelength of the excitation light is 700 nm; the low-index side of the
metal layer consists of water (e3 ¼ 1.77).
remarkable sensitivity that the SPR condition has for a changing dielectric
environment.
Phase jump. As already mentioned in Section 2.3.1 and expressed in eq. (2.6),
a reflection event at an interface is generally accompanied by a phase jump of
the reflected field. This is illustrated in Figure 2.5a for a prism–gold–water
system.
For comparison, the ‘‘conventional’’ SPR dip is shown in Figure 2.5b for the
same layer system. We see that around the SPR dip the phase of the reflected
electric field undergoes a relatively large change. The significance of this phe-
nomenon for sensing purposes is more clear when we plot the reflectance and
phase changes as a function of incident angle for a certain change in dielectric
constant of the water. This is depicted in Figure 2.5c and d. In the following
rough calculation, we assume that both the change in reflection coefficient DR
and the phase change Dj are proportional to De. From Figure 2.5c we estimate
that DR/De E30, whereas from Figure 2.5d we find that Dj/De E250.
Experimentally, a minimum DRE103 can be measured, whereas a mini-
mum Dj E 103 is feasible, using interferometric techniques. The conclu-
sion is that on the basis of reflectance measurements a minimum DeE4 105
can be detected, whereas a phase measurement provides a sensitivity of
De E4 10–6. In view of the very approximate character of this calculation,
View Online
Figure 2.5 Comparison of the angle-dependent phase changes (a, c) and reflectance
changes (b, d) for variation of the dielectric constant at the low-index side
of the metal layer. A gold layer is used, SPs are excited at l ¼ 700 nm.
(c) and (d) depict the differential phase and reflectance, respectively, for a
change in the medium’s dielectric constant of 0.01.
the absolute values found are of limited validity; however, the finding that a
phase measurement provides an order of magnitude better sensitivity is a hard
conclusion and, indeed, this was demonstrated by Nikitin and co-workers [8,9].
The only drawback of this approach seems to be the much more complicated
experimental setup.
SP coherence length. Generally, the metal’s dielectric constant e2 is complex
and this circumstance results in a complex propagation constant
kx ¼ kx 0 0 + jkx 0 0 [cf. eq. (2.8)], where kx 0 0 and kx 0 0 are real and imaginary parts,
respectively. For a surface plasmon, traveling along the interface with wave-
vector kx, this implies that the field intensity decays with a characteristic
distance 1/2kx 0 0 . For gold and silver, the standard metals in sensor applications,
View Online
24 Chapter 2
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
Figure 2.6 SPR response to a dielectric step at several wavelengths. For each wave-
length the light angle of incidence is set such that outside the strip
(extending from 0 to 125 mm) the interrogating kx is resonant with the
surface plasmon wavevector. The surroundings of the strip has dielectric
constant e3 ¼ 1.
the imaginary part of the dielectric constant increases with decreasing wave-
length and the SP propagation length decreases accordingly.
This is illustrated in Figure 2.6: here a layer system was prepared where a
30 nm SiO2 strip was deposited on a 50 nm silver layer. For a series of wave-
lengths the angle of incidence was chosen such that SPs were excited in the area
View Online
edge of the strip the SPR condition is again fulfilled and the SP resonance builds
up. The figure nicely demonstrates that with decreasing wavelength the SP
propagation length becomes shorter: the blurring on the left side of the strip
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
becomes less prominent for shorter wavelengths. It turns out that in the wave-
length range 500–00 nm the propagation length varies between o10 and 40 mm.
For a quantitative description of the findings depicted in Figure 2.6, we have
to analyze the interference between the several fields that are present in the layer.
In Figure 2.7, the layer system and the fields involved are indicated: the
resonant SP field, the non-resonant SP field and the external exciting field with
amplitudes E1, E2, E3, respectively. For a resonant SP that enters the SiO2 strip-
covered layer, the total field reaching the detector can be written as [10,11]
0 00 0 0
j k1 þjk1 x
Etot ðxÞ ¼ ðE1 E2 Þe þ E2 ejk0 x E3 ejk0 x ð2:11Þ
where k 0 1 is the wavevector corresponding to resonance in the covered area and
k 0 0 is the wavevector that excites SPs in the uncovered area.
Defining A ¼ E1E2 and B ¼ E2E3, the resulting intensity at the detector
becomes
00 00 0 0
Itot ðxÞ ¼ B2 þ A2 e2k1 x þ 2ABek1 x cos k1 k0 x ð2:12Þ
When a non-resonant SP leaves the covered area, the resonant SP builds up and
the intensity at the detector decreases accordingly [11]:
h 00 i2
Itot ðxÞ ¼ B þ A 1 ek0 x ð2:13Þ
From Figure 2.6, we see that this model gives a very accurate description of the
experiments.
Figure 2.7 Definition of wave vectors and fields for the system consisting of a
dielectric strip on top of a metal layer.
View Online
26 Chapter 2
Both this model and the experiments indicate that a plasmon needs roughly
four times the propagation length Lx for a full decay or for a full build-up; this
propagation length can be loosely defined as
1
Lx ¼ ð2:14Þ
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
2k00x
This implies that SPs with mutual distances significantly larger than Lx are
independent. This is a very important conclusion because it is the fundament of
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
surface plasmon microscopy [12,13], with its many applications in SPR imaging
and SPR multisensing: on a substrate we can define areas that in an SPR
experiment will behave mutually independently, provided that these areas are
significantly larger than Lx2. For SPs on gold, excited at l ¼ 632 nm, LxE7 mm
and on a total sensor area of 1 cm2 more than 104 independent sensor ‘‘patches’’
that each have an area of somewhat smaller than 100 100 mm2 can in principle
be defined, of which the optical responses can be simultaneously read out by
using an imaging system. As practical aspects are outside the scope of this
chapter, the interested reader should consult Chapter 7 for more details.
Figure 2.8 The SPR ‘‘dip’’ for 46 nm of silver (dashed) and gold (solid) with water
on the low-index side, for several excitation wavelengths. Dielectric data
for the metal layer obtained from refs. [3] and [33].
28 Chapter 2
layers with dielectric constants and thicknesses ei and di, respectively, placed
between a prism with dielectric constant ep and a medium (e.g. water) with
dielectric constant ew (see. e.g., ref. [5]):
k ky;p ky;w
M11 þ M12 ey;w ep M 21 þ M 22 ew
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
w
rp ðaÞ ¼ ð2:15aÞ
ky;w ky;p k
M11 þ M12 ew ep þ M21 þ M22 ey;w w
M ¼ M1 M 2 . . . MN ð2:15bÞ
with
" jei #
cos ky;i di ky;i sin ky;i di
Mi ¼ jky;i ð2:15cÞ
ei sin ky;i di cos ky;i di
Of course, with the use of this equation. we face the same problems as those
when we use eq. (2.15).
It is seen that G can reach enormous values for e close to 2e1; in a ‘‘normal’’
dielectric medium where e1 4 0, this condition points to the use of a metal,
where e can be negative; additionally, the imaginary part of e should be as small
as possible. It turns out that this condition corresponds to the excitation of a
surface plasmon in the metallic nanoparticle [23]. Particularly in the field of
Raman spectroscopy this can result in enormous sensitivity enhancements7 (for
a review, see ref. [24]).
7
Conventional Raman spectroscopy suffers from a very low scattering efficiency which can be 12
orders of magnitude lower than that of fluorescence.
View Online
30 Chapter 2
Now consider a Raman-active molecule near a metal nanoparticle. The
detected Raman intensity I(o,osc) can be expressed as
2 2
IRaman ðo; osc Þ ¼ gEexc Esc ¼ gG2 ðoÞG2 ðosc ÞI0 ðoÞI0;sc ðosc Þ ð2:18Þ
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
where Eexc is the total excitation electric field to which the molecule is exposed
and Esc is the total Raman-scattered field. The constant g is an experimental
constant that is unimportant in the present discussion.
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
shape of the nanoparticles and computer programs are available in the public
domain that can predict the extinction spectrum of nanoparticles of any shape
[30], by modeling the particle as a series of dipoles placed in an oscillating
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
32 Chapter 2
aspects of society, varying from food safety monitoring and high-throughput
screening to early in vivo detection of tumor growth.
2.7 Questions
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
water
gold
2.8 Symbols
A A ¼ E1 E2
Aext extinction coefficient
B B ¼ E2 E3
const experimental constant in surface-enhanced Raman spectroscopy
c propagation velocity in vacuum
di film or layer thickness
E electric field strength
e charge of electron
E0 amplitude of electric field
E0 external applied field E0 (Section 2.5.2)
Eexc total excited electric field (Raman)
Ei incident electric field
Er reflected electric field
View Online
k wavevector
ky y component of wavevector
kx, net x component of wave vector
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
References
1. A. Otto, Z. Phys., 1968, 216, 398.
2. E. Kretschmann and H. Raether, Z. Naturforsch., 1968, 230, 2135.
3. H. Raether, Surface Plasmons on Smooth and Rough Surfaces and on
Gratings, Springer, Berlin, 1988.
4. M. Cardona, Am. J. Phys., 1971, 39, 1277.
View Online
34 Chapter 2
5. J.R. Reitz, F.J. Milford and R.W. Christy, Foundations of Electromagnetic
Theory, Addison-Wesley, New York, 1993.
6. D.C. Cullen, R.G. Brown, C.R. Lowe, Biosensors, 1987/88, 3, 211.
7. E. Hecht, Optics, Addison-Wesley, New York, 2002.
8. A.N. Grigorenko, P.I. Nikitin and A.V. Kabashin, Appl. Phys. Lett., 1999,
Published on 10 March 2008 on http://pubs.rsc.org | doi:10.1039/9781847558220-00015
75, 3917.
9. A.V. Kabashin and P.I. Nikitin, Opt. Commun., 1998, 150, 5.
10. B. Rothenhäusler, W. Knoll, J. Opt. Soc. Am. B, 1988, 5, 1401.
Downloaded by Syddansk Universitetsbibliotek on 06/10/2016 18:41:43.
11. C.E.H. Berger, R.P.H. Kooyman and J. Greve, Opt. Commun., 1999, 167,
183.
12. E. Yeatman and E. Ash, Electron. Lett., 1987, 23, 1091.
13. B. Rothenhäusler and W. Knoll, Nature, 1988, 332, 615.
14. B.P. Nelson, A.G. Frutos, J.M. Brockman and R.M. Corn, Anal. Chem.,
1999, 71, 3928.
15. S. Patskovsky, A.V. Kabashin, M. Meunier and J.H.T. Luong, J. Opt. Soc.
Am. A., 2003, 20, 1644.
16. K. Tiefentahler, W. Lukosz, J. Opt. Soc. Am. B, 1989, 6, 209.
17. R.G. Heideman, ‘‘Optical waveguide based evanescent field immunosensors’’,
PhD Thesis, University of Twente, Enschede, 1993.
18. T. Liebermann and W. Knoll, Colloids Surf. A, 2000, 171, 115.
19. H. Kogelnik, in ‘‘Topics in Applied Physics’’, vol 7, ed. T. Tamir, Springer,
Berlin, 1975.
20. R.W. Ditchburn, Light, Blackie, London, 1963.
21. S. Wang, S. Boussaad and N.J. Tao, Rev. Sci. Instrum., 2001, 72, 3055.
22. P.S. Vukusic, J.R. Sambles and J.D. Wright, J. Mater. Chem., 1992, 2,
1105.
23. U. Kreibig and M. Vollmer, Optical Properties of Metal Clusters, Springer,
Berlin, 1995.
24. K. Kneipp, H. Kneipp, I. Itzkan, R.R. Dasari and M.S. Feld, J. Phys.
Condensed Matter, 2002, 14, R597.
25. S. Nie and S.R. Emory, Science, 1997, 275, 1102.
26. K. Kneipp, Y. Wang, H. Kneipp, L.T. Perelman, I. Itzkan, R. Dasari and
M.S. Feld, Phys. Rev. Lett., 1997, 78, 1667.
27. C.R. Yonzon, D.A. Stuart, X. Zhang, A.D. McFarland, C.L. Haynes and
R.P. Van Duyne, Talanta, 2005, 67, 438.
28. C.L. Haynes, R.P. Van Duyne, J. Phys. Chem. B, 2001, 105, 5599.
29. S. Link, M.A. El-Sayed, J. Phys. Chem. B, 1999, 103, 8410.
30. http://www.astro.princeton.edu/Bdraine/DDSCAT.6.1.html
31. S. Schultz, D.R. Smith, J.J. Mock and D.A. Schultz, Proc. Natl. Acad. Sci.
USA, 2000, 97, 996.
32. A.J. Haes, D.A. Stuart, S. Nie and R.P. Van Duyne, J. Fluor., 2004, 14,
355.
33. K. Johansen, H. Arwin, I. Lundström and B. Liedberg, Rev. Sci. Instrum.,
2000, 71, 3530.