Ahmed Et Al (1994) - Effect of Multiple Light Scattering and Self-Absorption on the Fluorescence and Excitation Spectra of Dyes in Random Media

7/24/2019 Ahmed Et Al (1994) - Effect of Multiple Light Scattering and Self-Absorption on the Fluorescence and Excitation Sp

1/5

Effect

of multiple

light

scattering

and

self absorption

on

the fluorescence

and excitation

spectra of dyes in random media

S. A. Ahmed,

Zhi-Wei

Zang, K.

M. Yoo, M.

A. Ali, and

R. R. Alfano

The

absorption, fluorescence, and

excitation spectra

of a dye

in a highly scattering

random

medium were

studied

experimentally.

The intrinsic

absorption spectrum of

the dye does not

change in the

presence of

scatterers,

but the presence of scatterers

in the media

will change

the observed

fluorescence spectra.

The

observation

is accounted

for by the

change in the photon trajectory

path length

for the fluorescence

emission.

The fluorescence of objects

embedded in absorptive

random scattering

media has been studied

recently

and shown to

enhance the detection and imaging

of

such objects.

The use of fluorescence-imaging

tech-

niques of this type

has potential

application in

medi-

cal diagnostic

techniques.'

In these imaging tech-

niques

the object to be detected,

which has or is

made

to have

fluorescent

properties (e.g.,

by dye injection),

is illuminated and made to fluoresce. This lumines-

cence is

then selected for

detection and

imaging, while

the illuminating

light is

filtered out. The image

quality is

improved further

when the surrounding

scattering

medium

is made partially

absorbing

at the

luminescence

spectrum. Since the

luminescent ob-

ject to be

detected in a typical medical

application,

1

-

5

such

as a breast

tumor, is itself constituted

from a

highly

scattering

medium, the effects of scatter

on the

emitted

luminescence

spectra itself are

of interest to

the further development of these

techniques.

These

medical

applications

are particularly

important since

the distribution

of

the luminescent

spectrum itself

may relate to medical conditions to be diagnosed.

In this

paper we present

the results of experiments

that examine the impact

of internal multiple

scatter-

The authors

are with the

Department of

Electrical Engineering,

Photonic

Application Laboratory,

Institute for Ultrafast Spectros-

copy and

Lasers, The City College

and the Graduate

Center of the

City University

of New

York, New York,

New York 10031.

Received 30

January 1992; revised

manuscript received

4 Octo-

ber 1993.

0003-6935/94/132746-05$06.00/0.

o 1994 Optical Society of America.

ing within

a luminescent

body on the luminescence

spectrum

observed

at the

surface.

The

intrinsic

emission

spectrum

of

a luminescent

medium,

such

as an organic dye luminifor,

is typically

Stokes

shifted

with respect

to its

absorption

spectrum.

The

presence

of

internal multiple

scatterers

within

a

luminescent

body can

affect the

luminescence

ob-

served

at the

surface

of the

bodyin three

ways:

(1) Self-absorption.

The

luminescence

emitted

from

inside a

random medium

reaches

the surface by

a random-walk

diffusion

process.

The

mean

random

path

length

of the

luminescence before

it reaches

the

surface

depends on

the depth

of

the luminophor

inside the body

and the

photon-transport

mean

free

path

in the random

medium.

The

deeper the lumino-

phor and

the higher

the density of scatterers, the

longer

is the

photon

path length. Because

the

emis-

sion spectrum

of

a luminophor

overlaps

its absorp-

tion spectrum,

emitted

fluorescent

light

experiences

greater absorption

at shorter

wavelengths

as it propa-

gates

through

a luminescent

medium

in solution to

the surface.

This preferential absorption

at the

shorter

wavelengths

would

mean

that the emission

that

finally

emerges

at the surface

of a body

would

show

a spectrum

relatively

diminished

at the shorter

wavelength

end

compared

with the intrinsic

fluores-

cence

of the luminophor;

i.e., the fluorescence

spec-

trum observed at the surface

would effectively

appear

to be

red shifted toward

the longer

wavelengths.

Clearly

the greater

the depth within

the

body

from

which

he radiation

originates,and/or

the greater

the

scatterer

density, the greater

would be

this red shift.

(2) Wavelength dependence of scattering. Elastic

2746 APPLIED OPTICS / Vol.

33, No. 13 /

1 May 1994


2/5

(Rayleigh) scattering by random

particulate scatter-

ers will scatter the shorter wavelengths preferen-

tially,

which means that luminescence emitted from

inside the body at shorter wavelengths undergoes

more scattering

to reach the surface than that at

longer wavelengths. Thus the shorter wavelength

light effectively

undergoes a longer random-walk

path

and experiences

more absorption,

and hence

the

remaining

luminescence spectrum is effectively red

shifted when emerging from the surface. This wave-

length dependence can be observed in random

media

with a small scatterer

with a diameter of less than the

wavelength of light. This wavelength dependence

may be different in tissues. There the scatterers are

generally much

larger than the optical

wavelength;

however, generally it can still be expected that the

larger wavelengths

will be less scattered and hence

penetrate deeper.

(3) Energy-level distortions. At sufficiently high

concentrations, scatterer densities might conceivably

be sufficiently high that

interatomic forces cause

distortion effects on the energy levels of the lumino-

phors, with impacts on their

intrinsic fluorescence

emission

spectra resulting.

The coherent

interfer-

ence between the radiated field of a dye molecule and

the field radiated from an effective Hertzian

dipole

induced on the closest scatterers may also conceivably

affect the emissions

spectrum.

6

This

effect has been

observed when more than one scatterer within a unit

wavelength exists. Such effects, if they exist, would

depend on the density of scatterers, as effects (1) and

(2) do. However, unlike the latter they might just as

well cause a blue shift as a red shift. Furthermore

their effects would not depend on the depth of the

emitting body within the random medium, as (1) and

(2) do.

To investigate the impact of multiple light scatter-

ing in a luminescent random media on the emitted

spectra, we carried out experiments on organic dye

luminophors

dissolved in liquid solvents with TiO

2

particles added to provide random scattering. Rho-

damine 6G (Rh6G) dye in 2-methyl-2,4-pentanediol

(MPD) solution was selected

as the luminophor, and

TiO

2

particles 0.18 plm n diameter were added to

provide scattering. This combination is similar to

that used previously to study backscatter image

enhancement in scattering media.'

We carried out fluorescence, absorption, and excita-

tion spectra measurements by using transparent

plastic cells of varying sizes containing mixtures of

the dye solution and TiO

2

beads. We investigated

fluorescence for two different cell arrangements with

a front surface and side excitation as shown in Figs.

1(a) and 1(b), respectively, by using a Perkin-Elmer

luminescence spectrophotometer LS50. The lamp

in the spectrophotometer was used as a light source

for excitation; the typical excitation wavelength was

514 nm.

We first obtained the intrinsic fluorescence of Rh6G,

excitation

detection

1cmI

1icm]

1

cm

(a)

excitation

detection

1

cm

(b)

Fig. 1. Excitation and detection

arrangement of dye fluorescence

in a transparent plastic cell.

in a 2 x 10-4 M concentration in MPD in a thin 0.016

cm x 1 cm cell by measuring front-surface excitation.

This measurement is

shown by the solid curve in Fig.

2(a), and it provides a standard for comparison. We

then examined the same concentration of dye solu-

tion in 1 cm x 1 cm cells using both front surface and

side excitation with and without the addition of the

known concentration of scatterers. In Fig. 2(a) we

show the results with the intrinsic fluorescence super-

imposed for comparison. The dashed curve in Fig.

^5~~~~

\

550 600 650 700

Wavelength )

(a)

550

600 650

700

Wavelength m)

(b)

Fig. 2. (a) Fluorescence spectra of 2 x 10-

4

-M Rh6G in a MPD

solution. Solid curve, front-surface detection from a 0.016-cm-

thick cell; dashed curve, front-surface detection from a 1-cm-thick

cell; dashed-dotted curve, side-surface detection from a 1-cm-thick

cell. (b) Fluorescence spectra of front-surface detection from a

1-cm-thick cell of 2 x 10-

7

-M Rh6G in MPD. Solid curve, dye

solution; dashed curve, dye solution plus TiO

2

scatterers.

1 May 1994 / Vol. 33, No. 13 / APPLIED OPTICS 2747

A I

I/l

\

II\

. '.

I .


3/5

2(a) shows

that

the front surface

fluorescence

of the

1-cm-thick

test cell

is red shifted

with

respect to the

intrinsic

fluorescence

(from the 0.016-cm-thick

cell).

This

red shift

is expected because

in the

thick cell

the

fluorescence inside

the

solution

contributes

to the

observed emission

spectrum.

Because

of

the longer

path the

fluorescence has

to propagate

through

the

dye solution,

the

shorter-wavelength

region of the

emission

spectrum will preferentially

be absorbed

by

the dye compared with the longer-wavelength region.

Thus the resultant

fluorescence

spectrum

from a

thick

cell is red

shifted when compared

with that

from a thinner cell.

This

effect is

further

enhanced

as shown

by the dotted-dashed

curve

where

an even

bigger red

shift results

with

the side

excitation since

the internally

excited fluorescence

now has

on aver-

age

an even longer

path

to the viewing

surface.

The

effect of a random-scattering

medium

on the

fluorescence

spectrum

of the

dye solution

is pre-

sented

in Fig.

2(b). The

solid curve in Fig.

2(b) is

the

fluorescence

spectrum

of 2 x

10-

7

-M

Rh6G in MPD

in

a 1-cm-thick

cell.

This

spectrum is itself

blue

shifted when compared with that for a considerably

higher

concentration

(3 orders of magnitude)

of dye

solution,

as shown by

the dashed

curve

in Fig.

2(a).

This observation

illustrates

the dependence

of the

fluorescence spectrum

on

dye concentration

and the

(expected)

effects

of self-absorption.

The addition

of

TiO

2

scatterers

into

the dye solution

results

in a red

shift

of the

luminescence

spectrum

obtained

with

front-surface

excitation

as shown in

Fig. 2(b)

by the

dashed

curve.

The

red shift

arises from the

longer

random-walk

path that the internally

excited fluores-

cence must now

undergo

in the

scattering

medium

to

reach the surface.

To examine whether the wavelength dependence of

particulate

scattering

plays

a part

in the

observed

effects, the

fluorescence

of

the test cells without

the

scatterers

was

viewed

through

a cell containing scat-

terers.

The

fluorescence

spectrum

transmitted

through the dye-free

scattering

medium

was

found to

be unaltered.

For

this purpose

the

experimental

studies

were

carried

out by the arrangement

in

Fig. 3.

In this arrangement

the

fluorescence

from

the 1-cm-

ArgonLaser

Fig. 3.

Schematic

of the experimental

setup for

studying

the

effect of the fluorescence spectrum when it passes through a dye

and a scattering

medium.

thick

dye

cell A that contains

2 x 10-

4

-M Rh6G

was

transmitted

through

cell B

without

and

with TiO

2

particles

suspended

in

MPD but with

no

dye.

The

fluorescence

spectrum

from

sample

A as

shown

by the

solid curve

in

Fig. 4

was found

to be essentially

unaltered

by the

presence

of (TiO

2

) scatterers

in

sample B

through

which

the

fluorescence

light was

permitted

to pass. Note

that

this

fluorescence

spec-

trum

is slightly

different

from

that shown

in Fig. 2(a),

because the emitted fluorescence spectrum is highly

dependent

on the experimental

geometry.

Using

a spectrophotometer,

we

also

measured a

direct

transmission

of the

optical

density

through

the

1-cm-thick

test

cell

that contained

a TiO

2

scattering

medium

without

a dye.

The optical

density

did

not

show

wavelength

dependence

from 550

to 650

nm.

Both

results

above indicate

that

the

wavelength

dependence

of light

scattering

in

the TiO

2

medium

is

a negligible

component

in the dye fluorescence

spec-

trum

observed

at

the surface,

and hence

it

is not

a

significant

factor in the

red shift

observed

in Fig.

2(b).

The tests

were

repeated

with

Rh6G

dye added

to cell

B, which contains the scatterers, and as expected the

red shift

is now observed.

This

red shift

is shown by

the

dashed

curve in Fig.

4, which

confirms

that the

red shift

in

the observed

fluorescence

spectrum

is

indeed

a result

of the

increased

random-walk

absorp-

tion

that

occurs

at

the shorter

wavelengths

of the

fluorescence

reaching the

surface.

To

examine

whether

the introduction

of

the TiO

2

scattering

beads

may have

affected

energy

levels

and

hence the

intrinsic

fluorescence

emission

spectrum

of

Rh6G,

a thin

0.016

cm x

1 cm test cell

was used

to

study

the

fluorescence

spectrum

in

the

presence of

TiO

2

scatterers.

A thin test cell

was used

so that

the

long trajectory path for the fluorescence light that

arises

from the excitation

of dye deep

inside

the cell

is

no

longer

a factor.

The intrinsic

fluorescence

spec-

trum of

Rh6G

(3 x 10-7 M)

obtained

by front-surface

excitation

fluorescence

measurements

is shown

by

the

solid curve in Fig.

5. With

the

addition

of TiO

2

scatterers

to the

solution

there

is a small red shift

a1

b

530

580

630

660

Wavelength nm)

Fig.

4. Fluorescence

spectra

(from

2 x

10-

4

-M

Rh6G in

a 1-cm-

thick cell)

when

it passes

through

a sample: solid curve,

with

no

scatterer

and with a TiO

2

scatterer

in the

MPD; dashed

curve,

TiO

2

in MPD

with a

Rh6G dye

(2 x

10-4 M).

2748

APPLIED OPTICS

/ Vol.

33, No.

13 / 1 May

1994


4/5

Wavelength )

Fig. 5. Comparison of the fluorescence of a thin (0.016-cm) cell

dye (3 x 10-

7

-M) with and without scatterers: solid curve, dye

fluorescence; dashed curve, dye with scatterers' fluorescence.

shown by the dashed curve in Fig. 5, which implies

that the larger red shifts observed in Fig. 2(b) in thick

cells (1 cm) are primarily a result of increased absorp-

tion at the shorter wavelengths and the increased

random-walk path length that the internally excited

fluorescence emission in the 1-cm-thick cell must

travel to reach the surface.

The effects of increased scatterer densities were

also examined in the 0.016 cm x 1 cm cells that were

used to obtain the results in Fig. 2(b). Using the

same dye concentrations (3

x 10-7 M), we gradually

increased the bead concentration and monitored the

front-surface excitation fluorescence for different con-

centrations. As the bead concentration was in-

creased, we observed that the red shift in the ob-

served fluorescence peak increased. A maximum

shift of the peak to as high as 7 nm was observed at

bead concentrations of 6.35 x 101

2

/cm

3

[Fig. 2(b)].

As the bead concentration was increased further, the

red shift was observed to decrease. At a bead concen-

tration of 1.? x 10

1

8/cm

3

the peak of the fluorescence

spectrum shown by the dashed curve (with scatterers)

is 4 nm blue shifted versus the solid curve, as shown

in Fig. 6. Both these effects may be understood

qualitatively in the following manner: As the scatter-

550 600 650 700

Wavelength n)

Fig. 6. Comparison of a dye (3 x 10-7 M) and dye with with

scatterers' fluorescence of a maximum blue shift in a 1-cm-thick

cell. Solid curve, dye fluorescence; dashed curve, dye with scat-

terer fluorescence with scatterers.

ing bead concentration is initially increased, the

internally emitted fluorescence undergoes an in-

crease in the length of the random walk it takes to get

to the surface, and hence it undergoes more absorp-

tion, which results in an increase in the red shift.

This effect predominates until the increase in the

concentration of scatterers reaches the point where it

starts to inhibit penetration of the excitation radia-

tion greatly, and hence the occurrence of excitation

and fluorescence emission are both restricted to

nearer the surface, with the length of the random

walks to the surface consequently shortened and

self-absorption and the red shift lessened. Thus two

competing effects are at work.

The excitation spectra of a dye solution are also

significantly effected by the presence of scatterers.

The curves in Fig. 7 show the excitation spectra of

DCM dye in MPD observed at 630 nm when side

excitation is used (Fig. 1). The observed excitation

spectra in the absence of TiO

2

scatterers is shown by

the solid curve, which shows two absorption bands at

300-420 and 530-590 nm, which contribute to emis-

sion at 630 nm. The contribution of 300-420-nm

absorption bands is significantly reduced in the pres-

ence of TiO

2

scatterers, which could be a result of

either increased scattering and/or absorption of UV

light by the TiO

2

scatterer, which reduces the amount

of excitation light that reaches the dye. In the

presence of scatterers the relative contribution of the

excitation at the 400-520-nm band increases as the

density of the scatterer increases. This observation

is attributed to the random-walk process of the

excitation light in the scattering medium. For the

less absorbing excitation light, such as the 400-

520-nm band, the scatter effectively increases the

length of photon travel in the medium and therefore

the interaction of the excitation light with the dye.

This process ? in turn increases absorption and

excitation in this relatively weakly absorbing 400-

3 v4,->X@--9^v~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~/ I

300

400

500

OO 700

Wavelength m)

Fig. 7. Comparison of the excitation spectra of DCM dye

(2.6 x 10-4 M) and dye with the scatterers increasing: solid

curve, dye fluorescence; dashed curve, dye with 0.005-g scatterers'

fluorescence; dashed-dotted curve, dye with 0.009-g scatterers'

fluorescence; dotted curve, dye with 0.039-g scatterers' fluores-

cence.

1 May 1994 / Vol. 33, No. 13 / APPLIED OPTICS 2749


5/5

520-nm spectrum and therefore increases the ob-

served fluorescence.

Figure 7 also

shows that

the

excitation

in the

520-590-nm band is reduced

where

the dye absorption

is stronger. This observation

shows that the

relative amount

of absorption for

a

stronger absorbing band is reduced with respect to

the weaker absorbing band

in a highly scattering

medium, because the weaker

absorbing light needs

the

longer random path

to enhance the absorption.

In conclusion, it is shown that the presence of

random scattering

in luminescent

bodies affects

the

spectrum of the fluorescent

radiation observed

at the

surface of

these bodies.

The primary

effects ob-

served are

red shifts in the emission

spectra with

respect to the intrinsic fluorescence emission. These

shifts may

be understood

in terms

of an increased

random walk and hence an increase in absorption

that emission internally

excited within the lumines-

cent body undergoes

when

random-scattering

pro-

cesses are introduced.

References

1.

K. M. Yoo, Z. W.

Zang, S. A. Ahmed, and R. R. Alfano, Imagine

objects hidden in scattering media using a fluorescence-

absorption technique, Opt. Lett. 16, 1252-1254 (1991).

2. D. B. Tata, M. Foresti,

J. Cordero, P. Tomashefsky, M.

A.

Alfano, and R. R. Alfano, Fluorescence polarization spectros-

copy and time-resolved fluorescence kinetics native cancerous

and normal rat kidney tissues, Biophys. J. Biophys. Soc. 50,

463-469 (1986).

3. Y. L. Yang, Y. M. Ye, R. M. Li, Y. F. Li, and P. Z. Ma,

Characteristic autofluorescence for cancer diagnosis and its

origin, Laser Surg. Med. 7, 528-533 (1987).

4. S. Montan and L. G. Stroemblad, Spectral characterization of

brain tumors utilizing laser-induced fluorescence, Lasers Life

Sci. 1, 275-285 (1987).

5. B. L. Horecker, The absorption spectra of hemoglobin and its

derivatives in the visible and near infrared regions, J. Biol.

Chem. 148, 173-183 (1943).

6. J. Martorell and N. M. Lawandy, Spontaneous emission in a

disordered dielectric medium, Phys. Rev. Lett. 66, 887-890

(1991).

2750 APPLIED OPTICS / Vol. 33, No. 13 / 1 May 1994

Documents

Ahmed Et Al (1994) - Effect of Multiple Light Scattering and Self-Absorption on the Fluorescence and Excitation Spectra of Dyes in Random Media