Title: Atomic fluorescence flame spectrometry with a continuous wave dye laser
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00097493/00001
 Material Information
Title: Atomic fluorescence flame spectrometry with a continuous wave dye laser
Physical Description: v, 64 leaves : ill. ; 28 cm.
Language: English
Creator: Smith, Benjamin Willard, 1951-
Copyright Date: 1977
Subject: Sodium -- Analysis   ( lcsh )
Fluorescence spectroscopy   ( lcsh )
Chemistry thesis Ph. D
Dissertations, Academic -- Chemistry -- UF
Genre: bibliography   ( marcgt )
non-fiction   ( marcgt )
Statement of Responsibility: by Benjamin W. Smith.
Thesis: Thesis--University of Florida.
Bibliography: Bibliography: leaves 61-63.
Additional Physical Form: Also available on World Wide Web
General Note: Typescript.
General Note: Vita.
 Record Information
Bibliographic ID: UF00097493
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: alephbibnum - 000185658
oclc - 03349504
notis - AAV2244


This item has the following downloads:

atomicfluorescen00smitrich ( PDF )

Full Text








The author wishes to express his sincere appreciation

to all of the members of the research group of Dr. J. D.

Wincfordner for much helpful discussion and assistance. In

particular a special expression of gratitude must be made to

Drs. David J. Johnson and Nicolo Omenetto for their experi-

mental assistance, theoretical insights and general encourage-


Dr. James D. Winefordner has been a remarkable research

director .and deserves much gratitude- for making the -author's

graduate education very rewarding.







Theory 5
Population Equalization Approach 7
Line Broadening Approach 12
Experimental Verification of Saturation 16
Experimental Procedure 19
Experimental Concentration Measurement
Under Conditions of Saturation 28
Experimental Procedure 31
Measurement of (BF) 31
Comparison Measurements of nT 35
Sodium Concentration Profiles for Sever-
al Flames 39
Conclusions 44

Experimental Procedure 46
Operation of the cw Dye Laser 47
Choice of Analytical Lines 51
Analytical Results 54
Discussion 58




Abstract of Dissertation Presented to the
Graduate Council of the University of Flocida
in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy



Benjamin W. Smith

March 1977

Chairman: James D. Winefordner
Major Department: Chemistry

The unable continuous wave (cw) dye laser has been shown

to be capable of producing saturation conditions for sodium

in an air acetylene flame. The experimental evidetncr indi-

cates that the sodium double is a reasonably good approxima-

tion of a two-level system. Saturation of the sodium 589.6 rn:

line was observed at laser spectral irradiances of 1013 erg
-1 -2 -1
s cm nm and greater. Measurements of the absolute

maximum fluorescence radiance under saturation conditions have

been used to measure sodium concentration profiles in air -

hydrogen and air acetylene flames with spatial resolution

on the order of 0.01 cm. The cw dye laser has also been used

as an atomic fluorescence exciration source for analytical

purposes and, using resonance fluorescence, limits of detection

of 0.0001 pg ml-I and 0.04 pg ml-1 have been obtained for

sodium and barium, respectively. Using nonresonance fluores-

cence with excited state transitions, limits of detection of

between 0.3 ig mi]- and 100 pg ml-l were obtained for Cu, Li,

Mo, Nd, Rh, Sc, V, and Sr. All transitions were between 550

and 620 nm using rhodamine 6G and sodium fluorescein as laser



Soon after the initial analytical development of atomic

flucresncnce flame spectrometry (AFFS) in 1964 (1), it was

realized that the dependence of the atomic fluorescence signal

upon the excitation source intensity was a primary limitation

to the available fluorescence signal. At that time, only con-

ventional incoherent excitation sources such as hollow cathode

lamps and microwave excited electrodeless discharge tubes were

avail-ble, and these suffered from the disadvantages of low

intensity and/or poor stability. Therefore, much of the ana-

lytical atomic fluorescence research in the ensuing yvers

was devoted to development of intense, stable sources of ex-

citation. The commercial appearance of the tunable dye laser

in the late 1960's has generated much interest in its potential

as the ultimate excitation-source for atomic fluorescence. The

physics of the dye laser has been reviewed by Shank (2) and

Webb (3).

The first application of the dye laser to AFFS appeared

in 1971 (4) when Fraser and Winefordner used a nitrogen pumped

dye laser to excite atomic fluorescence for nine different

elements in hydrogen-air and acetylene--air flames. They re-

ported good limits of detection (within 10- or 100-fold of the

best reported literature values obtained with conventional

sources) and linear dynamic ranges of about four decades.

Their system consisted of a nitrogen laser pumped organic dye

laser with peak power of 10 kW, a pulse width of 6 ns, a spec-

tral bandwidth of 1 nm, and a repetition frequency of 1-30 Hz

depending upon the dye used. Signal detection and processing

were carried out with a 1P28A photomultiplier tube with fast

response dynode chain and a boxcar integrator with a 10 ns

gate width.

In the same year, Denton and Malmstadt (5) studied the

atomic fluorescence of barium excited by a Q-switched ruby

laser pumped dye laser and found that an ultrasonic nebuliza-

tion system was helpful in reducing scatter.

In 1972, Fraser and Winefordner (6) obtained good results

for resonance AFFS, excited state resonance fluorescence,

Stokes direct line fluorescence, anti-Stokes direct line

fluorescence and excited state anti-Stokes direct line fluo-

rescence with the same nitrogen laser pumped dye laser system

described above. They utilized nonresonance fluorescence to

obtain a 10-fold improvement in limits of detection due to

decreased scatter. Results such as these point out the versa-

tility of the tunable dye laser as a variable wavelength

excitation source. Omenetto et al. (7) studied the laser ex-

cited atomic fluorescence of transition elements in the

nitrous oxide-acetylene flame in 1973, again using a nitrogen

laser pumped dye laser. They stressed the value of nonreso-

nance fluorescence as a means of minimizing spectral inter-

ferences and reducing scatter. Omenetto et al. (8) also

examined the atomic and ionic fluorescence of the rare earth

elements with a nitrogen pumped dye laser system, observing

mainly nonresonance fluorescence transitions; they obtained

limits of detection equal or close to the best atomic absorp-

tion techniques.

The tunable dye laser has several advantages over con-

ventional line sources: only one source is needed and the

source radiation is coherent and thus easily focused to ob-

tain large power densities. With low duty cycle pulsed laser

systems, such as were used by the above authors, the advantage

of reduced flame background noise was expected, but not gener-

ally realized for various reasons (9). Up to this point in

time, no consideration had been given by workers in the field

of laser excited AFFS to the possible influence of extremely

high source intensities upon the atomic population conditions

within the flame. Omenetto et al. (10) and Piepmeier (11)

first derived the theory necessary to visualize the influence

of saturation of atomic levels by intense laser excitation. Kuhl,

Neumann and Kriese have discussed the influence of saturation

on flashlamp pumped dye laser excited AFFS (12).

During excitation by conventional low intensity light

sources, the fluorescence irradiance is known to be linear

with source intensity (13). This relationship holds only up

to the point at which the intensity of the source becomes

large enough to alter the thermal atomic distribution in the

flame. A nearly saturated fluorescing atomic population in

a typical analytical flame results in emission signals orders

of magnitude larger than the thermally excited emission.

Furthermore, under conditions of saturation, the fluorescence

is no longer linearly dependent upon source intensity; it may

follow a squareroot dependence or even be independent of

source intensity, depending on the type of laser source em-

ployed. Also, at saturation the fluorescent emission shows

little dependence on changes in collisional quenching (e.g.,

quantum efficiency) while retaining a linear dependence upon

atomic vapor density within the hot gases (assuming the vapor

is dilute).


The attainment of saturation within an atomic vapor

provides a means of measurement of absolute concentrations

of atomic species in an analytical flame. In this section

will be given a detailed theoretical treatment of steady state

saturation by two different approaches, a verification of

the effect experimentally and its application to measurements

of concentration profiles in several flames of analytical



There are two distinct approaches available for a theo-

retical evaluation of the saturation phenomena. The first,

and most straightforward, follows the derivation given by

Omenetto and Winefordner (14); it is based on the consideration

that the absorption coefficient goes to zero at high excitation

irradiance because the populations of the levels involved are

equalized. In the saturated steady state, the rate of absorp-

tion and the rate of emission are equal and no net absorption

may occur. This will be referred to as the "population equali-

zation" approach.

A second method for arriving at an expression for fluo-

rescence irradiance at high source intensity is suggested by

Killinger et al. (15) in their derivation for laser saturated

gas phase molecular fluorescence of OH and by Pantell and

Puthoff (16). An extension of their approach to the atomic

case leads to the same result as that given by Omenetto and

Winefordner. Because the derivation is based upon a considera-

tion of linebroadening as the controlling parameter affecting

saturation, this will be called the "broadening" approach.

For either derivation, the following assumptions are

made: (i) the atomic system is characterized as an ensemble

of atoms having only two energy levels 1 and 2 without multi-

plet splitting, with equal statistical weights gl = g2, sepa-

raLed by the energy difference E = hvo, and having populations

n, and n2, with n1 + n2 = nT, where nT is the total density

of the atoms; (ii) the atoms are present in thermodynamic

equilibrium at low concentration (i.e., selfabsorption is not

considered); (iii) the laser radiation does not affect the

energy distribution of the gas molecules, the velocity dis-

tribution of the atoms, and the temperature T of the system;

(iv) polarization effects are neglected; (v) the atomic system

is spatially homogeneous with respect to both concentration

and temperature; (vi) the radiation density of the source is

spatially homogeneous and constant while traversing the system;

and finally (vii) steady state conditions have been achieved.

It is also necessary to make some assumptions concerning the

spectral nature of the laser source with respect to the atomic

absorption profile of the atoms within the flame. The situation

is complicated by the fact that the spectral output of the

continuous wave (cw) dye laser is not uniform but consists

of a series of narrow spectral modes within the narrow spec-

tral output produced by the tuning element within the dye

laser cavity. For the laser used in this study, the overall

output linewidth is 0.08 nm (FWHM). This linewidth is com-

posed of some 65 spectral modes, each mode having a width of

8x106 nm with a space between modes of 4x104 nm. Therefore,

over a typical atomic absorption linewidth in the flame of

0.01 rim there are about 25 individual spectral components
each about 10- nm wide. The interaction of each of these

modes with the absorption profile causes considerable cclli-

sional and Doppler broadening resulting in a more or less

continuous absorption over the atomic absorption profile.

The assumption is made, then, that the laser acts as a quasi-

continuum source, i.e., the overall linewidth of rhe laser

is sufficiently broader than the atomic absorption width

in the flame such that the source may be considered to approach

a continuum. The experimental results appear to confirm this


Population Equalization Approach

Based upon simple kinetic considerations (14), the time

dependent concentration of the excited state, n2, can be rep-

resented as

d- = Bpv (n1 n2) + kl2n n2(A12 + k21) (1)

where p is the uniform spectral volume energy density at

S v (erg Hz cm -3) of the quasicontinuum laser, B (erg-

cm3 s-1 Hz) is the Einstein coefficient for absorption and

stimulated emission (B12 B21 = B, since gl = g2), k12 (s-)

is the unimolecular rate constant for thermal excitation, k21

(s-1) is the unimolecular rate constant for thermal or non-

radiational de-excitation, A21 (s-) is the Einstein transition

probability for spontaneous mission and n1 and n2 represent

the atomic concentrations (cm ) residing in levels 1 and 2.

At the temperatures of typical analytical flames (2000-4000 K)

and for typical energy differences of 2 eV or less, kl2 is
negligible with respect to k21; and since T = (A21 + k21)-

equation 1 reduces to

dn2 n2
dt Bp nl Bp 2 (2)
o 0

where T (s) is the mean radiative lifetime of level 2. Since

n1 = nT n2

dn2 n2
dt+ (nT n2) Bp n2
o o


dn2 n2
dt + T Bp nT 2Bp n2


dn2 n2
-d- + (2Bp 1) Bp n (3)
o O

From the theory of radiation

B (nI n2) / k(v) dv (4)
o line
-1 m-l
where c is the velocity of light (cm s-) and k(v) (cm-) is

the frequency-dependent atomic absorption coefficient of the

atomic vapor. The frequency dependent absorption cross-sec-

tion o(v) (cm2) is related to k(v) by (17)

fo(v) dv = (nl n2)-1 f k(v) dv (5)

which converts equation 4 to

B = T- o(v) dv (6)
o line

Substituting equation 6 into equation 3 gives

dn2 n2 2c
d + T2 hv Pv f o(v)) dv H 1]
So line

c p T I o(v) dv (7)
o a line

For convenience and simplification let E = p c, where
o o
-1 -2 -
E (erg s cm- Hz ) is the spectral irradiance of the laser.
By defining a saturation spectral irradiance parameter Es
as (18)

S= (2T fo(m) dv)- (8)
ihvon 7 bo

equation 7 becomes


dn2 n2 E
2 [ + 1- =
dt Es


hvo T / o(v) dv (9)
o line

For the case of a dilute atomic vapor, the fluorescence
radiance, DF (erg s cm2 sr ) is given by (13,19)

B = n2 hvo (rsp)- (10)

where 4 (cm) denotes the depth of the homogeneous fluorescing

volume in che direction of observation and Tsp (s) is the

radiative lifetime of level 2.

By substituting into equation 9 and rearranging terms

E v o nT
dB B- v v
S[1 + ] 0 f av) dv (11)
dt Es sp line

-1 -1

or, if a constant C (Hz sr s ) is defined as

1 nT
C 1 2 T f oc(v) dv
sp line


[1 + -- = E C
dt T Es vo
For a continuous wave dye laser, steady state conditions are

assumed, i.e., dBF/dt = 0; thus
E +- Es
BF = CE [ o -r[ (12)
vo E

In the case of a pulsed laser, steady state conditions may

also hold provided that the pulse width of the laser is

greater than .

Reintroducing the terms involved in the constant C

B z E T n a(v) dv (13)
F 4-[ o Tsp E + E line
o V

Remembering that the transition probability is defined as

A21 s ( -)1 and substituting equation 8

hv V
BF = E A21 --2 n, [ ] (14)
oF 21 E i E + ES

Now, if the source irradiance E >> Es (high irradiance case)
Vo V
o "o
saturaLion has occurred and equation 14 reduces to

(BF) max 2 h T (15)
()a = a -f A21 ho -15)

where (BF)max represents the maximal fluorescence radiance

obtainable for a given nT. Thus, an absolute measurement of

(BF)max yields the product of nTA21, all other parameters being

known. The great advantage of determining this product under

conditions of saturation is that the measurement is independent

of source spectral irradiance, temperature of the atomic

system, and quantum efficiency for the transition.

The physical significance of the saturation parameter is

easily seen if one lets E = E in equation 14. In this
0 0
case BF = (Br) a/2. Thus, Es represents the source spectral
F max Vo
irradiance for which the maximum attainable fluorescence radi-

ance is decreased by a factor of 2.

Furthermore, if E << E (low irradiance case) equation
o 'o
O 0
14 becomes

A Vo nT
B = s A21 hvo (16)

and since, at low irradiance nT n1 and by reintroducing

the definition of E remembering equation 5 and the defini-
tion of quantum efficiency, Y21= "/sp

B = 21 E f k(v) dv (17)
o line

Equation 16 shows the well-known linear dependence of the

fluorescence radiance upon the source spectral irradiance

and the quantum efficiency of the transition.

Line Broadening Approach

Equations 15 and 17 may also be obtained by a derivation

based upon line broadening under the influence of a high radi-

ation field (laser beam) (15). The same assumptions given

above are required. When the source spectral width is greater

than the atomic absorption width within the flame, the rate

of absorption, Rabs (photons s 1) is given by

Rabs = S f EpSv{1 exp[-k(v)Z]} dv =


S f k(v) EpSv dv


S = cross-section of exciting beam (cm2);

EpSv = source spectral photon irradiance at frequency
v (cm-2 s- Hz-1);

k(v) = absorption coefficient at frequency v (cm-);

9, = absorption path length (cm);

and wheh the expression on the right is evaluated for low

optical density, i.e., k(v) k Z 0.05, equation 18 can be re-

written in terms of nT (cm 3) and the absorption cross sec-

tion o(v) (cm2) as

Rabs = S nTl ePSSv C(v) dv (19)

The absorption cross-section depends upon the broadening

processes and the approach to saturation with high intensity

sources, the absorption cross-section being the ratio of the

power absorbed per absorbing species to the incident power

per unit area, i.e., the effective area per absorber which

removes energy from the incident beam.

For the case of pure homogeneous (Lorentzian) line

broadening and for continuum excitation

Eps a o 1
SPSva o(-) dv = (-- (- (20)
line s s

where EpS (cm-2 s-) is the integrated source irradiance,

o0 (cm2 Hz) is the integral of o(v) over the line, 6vs (Hz)

is the source spectral width (FWHM) and Ss is defined as

S 2- 2v
s o 6v



Thus, the rate of absorption is given by

EpS Co 1
S "T S -
T 6V s +- -- s)

Equation 22 may be rewritten

Rabs S= S nT o(v) dv C- n- )
line s o c

where EpS = Ec/hv\' and E is the source spectral
-1 -2
in erg s cm

Making use of the previous definition of the

irradiance, ES (equation 8), equation 23 becomes
i n


hv E
S k n 0 o
n 2Es T hv s (I- +T s







Now, after


S nT Ec
26v -


EPS to Ec


Es+ Es S
Vo O

of equation 8 into equation 21 and

S c
s Es 6v

and by substituting into equation 25




S nT EC 1
abs 2 6vs Es + (E /6v )


V 6,J
0 S


S T nT o
Rabs 2 (Es +s
E +E
0 0

Since the rate of fluorescence (photons s ) is given by (20)

Rabs Y21
Rfl 4-


S k nT Y21 Vo
fl 8 T Es + E

By substituting the definition of Y2 /Tsp, defining the

fluorescence radiance per cm2 and since each photon has an

energy of hv

nT A21 hv v
8F8 Es + E
O o

Multiplying through by ES /Es
o o

Z nT A21 hv E o0
BF = 2 ( ) (26)
41 2 Es Es + E
0 0 0

which is identical to equation 14. Substitution of the limiting

cases for high and low irradiances leads to the same results

because at low irradiance nT = n1. Although the above deriva-

tion assumes only homogeneous broadening, if the assumption of

pure inhomogeneous broadening is made, the same result is ob-

tained (15,16).

Experimental Verification of Saturation

Before discussing the experimental verification of

saturation, it is worthwhile to obtain some numerical concept

of the laser irradiance required to produce saturation as

well as the fluorescence radiance to be expected. The satu-

ration irradiance defined in equation 8 may be rewritten in

terms of variables which are more readily accessible by sub-

stituting from classical radiation theory (18)

fo(v) dv = (ie2/mc) f12 (27)

8 7T v e
( -sp)1 ic f12 (28)

where e and m are the charge and mass of the electron and f12

is the classical absorption oscillator strength. Therefore

4Ti hv 3 T 4n hv
Es --2 () = 0 2 (29)
0 C c

or, converting frequencies to wavelength and evaluating all


Es = 7.6 x 1023 5 Y1 (30)
A o 21

where A has units of nm, Y21 is dimensionless and Es has
-2 -1 -1
units of erg cm nm s For the sodium line at 589.6 nm
11 -2 -1 -1
the saturation irradiance is 3.6 x 10 erg cm nm s,

assuming a quantum efficiency of 0.03 which is both a rea-

sonable estimate and a literature value (21) for an air-

acetylene flame very similar in composition to the one used

in the following experiments. The shape of the BF vs. E.
curve may be obtained from equation 14 for a given value of

n. since the transition probability for the sodium double is

well known (22). Figure 1 presents a log-log plot of this

curve for several values of Y21. For a quantum efficiency

of 1.0, the saturation irradiance is 1.1 x 010 erg s- cm2

nm1 which must represent rhe very minimum source irradiance

required to approach saturation. In plotting the curves, nT

is assumed to be 1010 cm3 and A = 100 pm = 0.01 cm. In all

calculations for sodium (589.6 nm) A21 is taken to be 6.28 x

107 s-1 (22).

The cw dye laser used in this study provides a maximum

output at 589.6 nm of about 1 W with a spectral width (FWIHM)

of 0.08 nm. Beyond this, the available spectral irradiance

at the flame depends only on the tightness with which the

laser beam may be focused. Because the output is TEM mode,

the focused beam waist diameter is given by (23)


; c2if

= .01

i 1 12 13 I
LOS(E ) rg 'l Scm' nmr"

Figure 1. Variation of Atomic Fluorescence with Source Spectral Irradiance for Sodium.

D 4 f (31)
Waist nd (31)

where X is the wavelength of light passing through the focussing

lens, f is the lens focal length, and d is the diameter of

the laser beam as it enters the focussing lens. In all cases,

a 3.5 in focal length lens has been used (f = 8.89 cm), and

the beam diameter is approximately 0.5 cm (measured photograph-

ically) as it enters the lens; therefore, the waist diameter
is 1.3 x 10-3 cm. This corresponds to a focused spot of

1.3 x 10 cm Therefore, the maximum available spectral

irradiance is given by

S1.0 W
E 2
o (0.08 nm)(1.3 x 10-6 cm2

06 2 -1
= 9.6 x 10 W cm2 nm-
13 -1 -2 -1
which corresponds to 9.6 x 101 erg s- cm2 nm and should

thus be nearly sufficient to achieve saturation in sodium if

Y21 0.01.

Experimental Procedure

A block diagram of the experimental system is shown in

Figure 2. The cw dye laser is pumped by all lines of a 4 W

argon ion laser. The dye laser output is beam-split to a

compact low resolution monochromator for monitoring the laser

power continuously via a photomultiplier and an associated

DC readout device. This laser energy monitor was calibrated

against a certified laser pyrometer at frequent intensity

intervals. A summary of the calibration is given in Figure 3

Laser Dye Las: r _
i n cr

J -----4--- I-----

Rf -_ _

Rccoroer Lo1 In Crent AImp
I _______ J A^ i;;lr i ________
______ _

Figure 2. Block Diagram of the Experimental System

Laser Power
Laser Power

Figure 3. Response of Laser Power Monitor vs. Laser Power

-- ----- ,-. ----- -------- 1


as a plot of:relative signal from the beam splitter readout

vs. measured laser power in milliwatts. The log-log plot

has a slope of 0.97. After passing through the chopper (260

Hz) the beam is steered by two mirrors (not shown) to the

height of the burner and focused by L1 (focal length 3.5 in)

onto the burner center. The collection optics consisted of

a single.8 in focal length lens at right angles to the laser

beam adjusted to produce a 1:1 image of the focused beam

onto the detection monochromator slit. The detection mono-

chromator was a 0.5 m focal length Czerny-Turner grating

spectrometer. The photomultiplier current is converted to a

voltage and synchronously detected by a lock-in amplifier

which is synchronized with the chopper. The lock-in output

is displayed en a strip chart recorder. Table 1 gives the

details of all experimental components. By placing neutral

density filters between the dye laser and the quartz plate

beam splitter, the laser irradiance at the flame may be accu-

rately varied over several decades. When necessary, a tungsten

strip lamp was placed at the position of the burner for absolute

calibration of the detection spectrometer. The calibration

procedure will be discussed later. Flame gases were all

standard commercial grades with no filtering. After two-

stage regulation at the cylinders the gases were passed through

calibrated rotometer flow meters fitted with needle valves

and sent to the premix chamber via Tygon tubing. Some fluc-

tuation in the acetylene pressure had been detected in previous

studies so a 10 A stainless steel ballast tank was placed in

Table 1
Experimental Components and

Model No.



Argon Ion Laser

Cw Dye Laser

Tungsten Strip




H. V. Power Supply

H. V. Power Supply


Strip Chart

Voltage Converter

Lock-in Amplifier

Signal Generator

Laser Power Meter

Strip Lamp Power



EP UV 1068
EP UV 1104







Servoriter II





Control Laser Corp.,
Orlando, FL

Coherent Radiation,
Palo Alto, CA

Eppley Laboratory, Inc.,
Newport, RI

RCA, Electronics Com-
ponents, Harrison, NJ

Hamamatsu Corp.,
Middlesex, NJ

Princeton Applied Re-
search, Princeton, NJ

Keithley Instruments,
Cleveland, OH

Heath Instruments,
Benton Harbor, MI

Spex Industries,
Metuchen, NJ

Texas Instruments Corp.,
Houston, TX

Keithley Instruments,
Cleveland, OH

Keithley Instruments,
Cleveland, OH

Wavetek, San Diego, CA

Coherent Radiation,
Palo Alto, CA

Laboratory Constructed,
0 50 A, 12 V DC,
University of Florida



Table 1 continued


Flow Meters

Burner Nebulizer

Front Surface

Model No.





Air Products Corp.,
Allentown, PA

Perkin-Elmer, Norwalk,

Melles Griot, Irvine,

the fuel gas flow line between the supply cylinder and the

flowmeters. A two-way valve was provided in the oxidant line

to select nitrous oxide or air.

In order to experimentally verify equation 14, it was

necessary to obtain values of relative fluorescence vs. laser

spectral irradiance over as wide a range of laser intensity

as possible. However, with the laser tightly focused and

with a 100 pm slit width to restrict the detection region

to the very center of the laser focus waist, it was found

that a severe reduction in signal-to-noise ratio occurred at

laser irradiances of more than two decades below the maxi-

mum. To achieve a reliable indication of the slope at low

irradiance it was necessary to remove the laser focussing

lens, thereby achieving a spectral irradiaLice reduction of

about five decades while maintaining an adequate signal-to-

noise ratio due to the much larger volume of sodium atoms

within the excitation region. Measurements were made, then,

over two regions of interest: a low irradiance case and a

high irradiance case. To establish the integrity of the

measuring system, data was taken on the Raleigh scatter re-

sulting from the tightly focused laser impinging upon a

flowing clean air stream. The log-log plot provided a slope

of 0.97 (3% low), indicating that the instrumental system

performed adequately.

Figure 4 shows the graphical summary of the results for

sodium saturation along with a theoretical curve for Y21
10 -3
0.03 and nT = 10 cm To satisfy the requirement of a

, ~~~:--3- (:,-...

/ Excrlenicl




r-1 S cm2 M
,"g s cm nra

Figure 4. Experimental and Theoretical
Source Spectral Irradiance

Curves for the Atomic Fluorescence of Sodium vs.

LC;S l (F)
(re 1 tlve)

Y =.03

dilute atomic vapor and to minimize any post- or pre-filter

effects, all data were taken while aspirating a 1.2 ppm

solution of sodium which provided about 1010 atoms cm-3 in

the flame gases. The results shown in Figure 4 were taken

on an air-acetylene flame with an argon shield. The ordinate

axis for the relative experimental data has been adjusted

to coincide with the theoretical curve.

Figure 4 clearly shows the effect of saturation at
13 -1
source spectral irradiances greater than 2 x 10 erg s
-2 -1
cm nm The average of the twelve points taken at the

top of the plot is 838 73 (relative units) which gives a
12 -1 -2 -1
saturation irradiance of 8.0 x 10 erg s cm nm From
equation 30 a quantum yield of 1.3 x 10-3 is calculated. The

fact that this quantum yield is about an order of magnitude

lower than the expected value of 0.03 (21) indicates that the

calculated value of saturation irradiance is probably too low.

The agreement between theory and experiment is nevertheless

satisfying when the uncertainties in the parameters involved

are considered and it is remembered that the simple model (14)

neglects all line profile effects and the influence of non-

quenching collisions. It is likely that efficient interlevel

mixing between the upper levels of the sodium doublet results

in a larger saturation irradiance than predicted since a
fraction of the laser excited sodium atoms in the 16956 cm

level may lose their absorbed energy by radiative decay via

the 16973 cm-1 level.

An attempt was also made to measure the effect of satura-

tion on the resonance barium transition at 553.5 nm using

fluorescein dye in the laser. However, the peak source spec-

tral irradiance of the dye laser at this wavelength is smaller
13 -1 -2 -1
by a factor of four (1.2 x 1013 erg s- cm2 nm ) and because

the saturation irradiance is proportional to -5, the barium
12 -1
saturation irradiance is calculated to be 1.5 x 101 erg s
-2 -1
cm nm1 (for Y21 = 0.01); therefore the laser was not in-

tense enough to induce a curvature in the plot of source

irradiance vs. relative fluorescence. The experimental data

with a slope of 0.97 are plotted in Figure 5.

Experimental Concentration Measurement
Under Conditions of Saturation

As previously mentioned, equation 15 provides a means

of obtaining absolute local species concentrations in a flame

by a measurement of the absolute spectral radiance of the

fluorescence signal (BF)max. Substituting the appropriate

parameters into equation 15 (see Table 2) gives for sodium

(589.6 nm)

(BF)max = 8.39 x 10-8 n (32)

where nT has units of cm3 and (BF)max is the maximum radi-

ance under saturation conditions (erg s-1 cm-2 sr-1). For

the verification of equation 32, the concentration nT was ob-

tained by two independent methods and compared with the value

obtained from a measurement of (BF)max

10 I1 2 -1 10
E. (erg s cm nm )

Figure 5. Experimental Variation of Atomic Fluorescence with
Source Spectral Irradiance for Barium

Table 2
Spectroscopic and Instrumental Parameters
for Sodium


Energy of the Transition (hvo)

Transition Probability (A21)

Path Length (W)

Statistical Weight (gl = g2)

589.6 nm

3.36 x 10-12 erg

0.628 x 108 s-1

0.01 cm


Experimental Procedure

Measurement of (BF)mm Because (BF)max must be known

in absolute radiance units, it was necessary to calibrate the

spectrometer detection system in absolute units. This was

done by placing a standard NBS-traceable tungsten filament

strip lamp at the flame so that the center of the tungsten

filament .corresponded to the point of minimum beam waist diam-

eter in the focused laser beam. A calibration factor was then

obtained by measuring the photocurrent corresponding to the

known spectral radiance of the standard lamp and accounting

for the difference in area of the filament and focused laser

beam and the spectrometer bandwidth as follows

iD A
C' = - (33)
(0 )aW ) B
Lamp fil

where C' (A erg- cm2 s sr) is the calibration factor, iD (A)

is the detector photocurrent, B (erg s' cm- nm sr )
is the absolute spectral radiance of the lamp at Xo, Alm (nm)

is the spectrometer spectral bandwidth and

A = (Dwaist)(S)


B = (h )(S)
where A (cm2) is the area of the focused laser beam passing

into the slit, S (cm) is the spectrometer slit width, B (cm )

is the area of the slit aperture (fully illuminated by the

tungsten filament) and hm (cm) is the slit height.


The calibration was made with two different lamps (see

Table 1) and the results coincided within 3%. The lamps

were operated at 35 A according to specifications of the

National Bureau of Standards. Remeasurement of the calibra-

tion factor on subsequent days (a total of four separate

measurements with each lamp) produced a repeatability of

about 5%. The values of the terms used in equation 33 are

summarized in Table 3. For 589.6 nm the value of C' was

2.0 x 10-12 (erg-1 A cm2 s sr).

A "standard" air-acetylene premixed flame was used for

all measurements of nT. The burner consisted of a commer-

cially available Perkin-Elmer premixed aspiration chamber

fitted with a laboratory constructed, stainless steel capillary

burner 1 cm in diameter. The burner head has been described

by Haraguchi and Winefordner (24). The capillary burner

head was surrounded by a laboratory constructed circular

casing through which argon was passed to act as an inert

sheath. Table 4 gives the pertinent burner characteristics.

A dilute solution of sodium in deionized water was prepared

for use in all of the following saturation-related experiments.

The sodium concentration was carefully determined by con-

ventional flame emission and found to be 1.2 ppm. The con-

centration was checked against freshly prepared standards

at several times during the course of the experiments and

found to be constant.

Aspiration of a 1.2 ppm sodium solution produced 5 x 10-

A of photocurrent corresponding to a (BF)max of 2.5 x 103

Table 3
Experimental Parameters for the Absolute Calibration
of the Instrumental System

Blp 6.50 x 10 erg s cm nm sr
(EP UV 1104, 589.6 nm)
(EP UV 1104, 589.6 nm)

5.78 x 10 erg s- cm-2 nm-I sr-1

(EP UV 1068, 589.6 nm)

1.3 x 10-3 cm


0.2 cm

0.01 cm

0.197 nm

Table 4
Flame and Nebulizer Characteristics
for the Standard Air-Acetylene Flame

Acetylene Flowa

Air Flowa

Aspiration Rate (q)a

Aspiration Efficiency (y)a

Solution Concentrationa

Total Gas Flow Rate (Q)a

Temperature (T)b

n T/1298c
Free Atom Fraction (B)d

Atomization Efficiency (c)c

aExperimentally measured
CFrom reference 25
From reference 26

2.60 t min-1

15.4 min-1

8.33 cm min-


1.2 pg ml-1
18 9 min1

2450 K




-1 --2 -1
erg s cm sr The laser spectral irradiance for this
13 -1 -2 -1
measurement was 4.6 x 10 erg s cm- n2m which, referring

to Figure 4, is sufficient to establish saturation. Then,
10 -3
from equation 32, nT = 3.0 x 100 cm3. The possible influence

of polarization of the laser beam on these measurements was

investigated and found negligible. Mathematically, the cal-

culation of (BF)max may be expressed as
(B 5 x 109 A -1 -2 -1
(BF'max C erg s cm sr

Comparison Measurements of nm. Independent measurements

of nT were carried out by two methods. First, the concentra-

tion of sodium atoms in the flame may be calculated from a

knowledge of the flame characteristics following the procedure

given by Winefordner et al. (25,27). This requires a knowledge
3 -1
of the total gas flow rate Q (cm s ), the solution aspira-
3 -1
tion rate, D, (cm min-1), the aspiration efficiency, y, (no

units), the flame temperature, T, (K), the free atom fraction,

8, (no units), the atomization efficiency, E, (no units), and

the concentration of the solution, Cm, (moles 11). The flainm

concentration is then given by

Cm n298 e6Y
n = 29 y (34)
S 3.3x10 "QTnt

where n298 and n. represent the number of moles of combustion

products present at room temperature and at flame temperature,

respectively. The calculation may be considered only as

approximate because several of the parameters (P, c, T) must

be estimated or obtained from literature sources. Neverthe-

less, it is useful to make the calculation and for the flame

specified in Table 4, nT = 6.1 x 101 cm-3.

An experimental value of nT may be obtained by a measure-

ment of the total absorption of the flame against a back-

ground continuum source at the wavelength of the transition

of interest. The procedure has been outlined by Zeegers and

Winefordner (28). The experimental setup is as previously

described except that a 600 W quartz iodine lamp is placed

behind the flame, collimated, chopped and focused at the

flame center. By scanning the spectrometer (with 0.03 nm

bandpass) across the sodium absorption line at either 589.6

nm or 558.9 nm, it is possible to obtain a value of the

total absorption, a, from which the flame concentration nT may

be calculated via

me Z
n uc z AX a (35)
T e 2 22' go m
where m (g) is the mass of the electron, c (cm s-1) is the

speed of light, e (cm3/2 1/2 s-1) is the charge of the elec-

tron, A (cm) is the peak wavelength of the absorption transi-

tion, f (no units) is the oscillator strength, A (cm) is the

path length through the flame, Z (no units) is the partition

function for sodium, AAm (cm) is the bandwidth of the spectrom-

eter, and a (no units) is the measured fraction of the incident

intensity absorbed. For the 589.6 nm sodium transition f =

0.655, go = 2, and Z = 2. Because I. = 1 cm, equation 35 re-

duces to

nT = 1.5 x 1012 a

and while aspirating a 1.2 ppm sodium solution a was found to

be 0.017. The total absorption method gives, then, a flame
10 -3
concentration of nT = 2.6 x 10 cm

Table 5 summarizes the results for the three determina-

tions of nT. The value obtained from equation 34 can be taken

only as a rough estimate, and so the discrepancy between it

and the saturation value is not disappointing. In particular,

the free atom fraction, 8, is in error by as much as 25% (26)

and the parameters which depend upon the gas flow rates (Q, 0)

are only known to within 10% (the reproducibility of the flow-

meters). Also, the flame temperature, T, has been taken from

the literature (26) and is, at best, accurate to 50 K for the

particular flame used in this study. These four variables

introduce a total uncertainty of about 60%. The value of

nT obtained by the total absorption measurement should be much

more accurate. Since all the variables in equation 35 are

well-known, the primary source of error in the total absorp-

tion result arises from the measurement of a. The signal-to-

noise ratio at a sodium concentration of 1.2 ppm was insuffi-

cient for an accurate measurement of total absorption, and it

was necessary to measure a at concentrations of 10 ppm and

higher and perform a graphical extrapolation to the lower con-

centration. Pronounced curvature was present in the growth

curve above 100 ppm, and so the extrapolation was carried out

on the linear portion of the curve from 10 to 100 ppm. The

plot obtained is quite linear (slope = 0.633, correlation coef-

Table 5
Results of the Three Determinations of nT


Total Absorptionb


3.0 x 1010 cm"3

10 -3
2.6 x 100 cm-3

6.1 x 1011 cm3

aFrom equation 32
From equation 35
CFrom equation 34

ficient = 0.997), and the good correlation coefficient indi-

cated that any error present probably amounts to less than 5%.

The 13% discrepancy between nT obtained via saturation and

the total absorption technique is then quite acceptable. The

largest source of error present in the saturation determina-

tion is the calculation of the laser beam focus waist which

is necessary in the calibration procedure. Equation 31 is

theoretical and assumes a perfect optical component with no

aberrations and a beam which is perfectly coherent. Because

neither of these assumptions is valid, there is a possible

error which would cause the beam waist to be larger than

calculated. This would increase the calibration factor, C,

(equation 33) leading to a smaller value of nT in the satura-

tion determination. Also, a larger beam waist would affect

the data shown in Figure 4 because the experimental values of

E, would be reduced, causing the experimental data to be

shifted to the left, i.e., a larger beam waist results in a

lower spectral irradiance.

Sodium Concentration Profiles for Several Flames

The use of laser saturation for the measurement of sodium

concentration profiles in flames has the distinct advantage

over more conventional techniques (such as total absorption

or absolute emission) of being a truly local technique. While

line-of-sight methods require an assumption of geometric sym-

metry within the flame gases and probe techniques tend to

disturb the point of measurement, the tightly focused laser

beam acts as a very high spatial resolution remote probe. This

superiority has been applied to two different flames in order

to evaluate its suitability as a local species probe.

The experimental setup is identical to that previously

described (Figure 1) except that the burner is mounted on a

two-dimensional translation stage enabling positioning in the

horizontal and vertical axis to within 0.05 cm. Profiles were

obtained by measuring the maximum fluorescence (BF)max at

various points throughout the flames. Table 6 lists pertinent

details of the flames studied. The hydrogen-entrained air

flame was chosen because it posesses a very cool inner zone

and provides an interesting horizontal profile. The air-acet-

ylene flame was studied because of its popularity as an analyt-

ical flame.

Figure 6 indicates the horizontal sodium concentration

profile in a hydrogen-entrained air flame at heights of 3.4 cm.

and 4.4 cm above the burner top. Horizontal profiles were

measured across the center of the flame. The aspirated solu-

tion concentration is 1.2 ppm. Figure 6 clearly shows that the

sodium profile is asymmetric and contains a concentration de-

pression of about 0.3 cm width at the flame center. The con-

centration is reduced at the higher profile which correlates

qualitatively with the vertical profile results in air-acet-


Figure 7 shows a vertical sodium concentration profile

for the air-acetylene flame. It is taken with the laser fo-

cussed at the flame center for an aspirated solution concentra-

Table 6
Flame and Burner Parameters for Flames Used
in the Profile Studies

Air/Acetylene Flame

Air Flow

Acetylene Flow

Solution Concentration

Hydrogen/Entrained Air Flame

Air Flow

Hydrogen Flow

Argon Flow

Solution Concentration

9.5 9 min-1

1.6 f mini

1.2 pg ml-1

7.7 9. min1
6.3 m i

1.2 pg ml-1

3.4 cm

.8 .4 .2 0 .2 .4 .6 .8 1.0
Distance from Burner Center (cm)

Figure 6. Horizontal Sodium Concentration Profiles in a
Hydrogen Air Flame at 3.4 and 4.4 cm Height
above the Burner Tip






0 2 4 6 3 10 12 14 16 18

Height above Burner Top (cm)

Figure 7. Vertical Sodium Concentration Profile in an Acetylene-Air
Flame Center

20 22

Flame at the

S- 9


tion of 1.2 ppm. There is a clear maximum in sodium concentra-

tion about 0.3 cm above the burner tip.


It has been shown that saturation of sodium atoms does

occur in flames under conditions of intense laser excitation.

Furthermore, the two-level steady state theory derived for the

process has been found to agree well with experiment. The

assumption that the 589.6 nm sodium transition acts as a simple

two-level system is certainly only approximate because the

doublet mixing cross-section is known to be quite large (O1015

- 1014 cm2 )(29). This probably accounts for the fact that

saturation was not achieved at as low of a source spectral

irradiance as was predicted. The potential of saturated

atomic fluorescence as a probe of atomic specie concentration

has been experimentally verified; and, where saturation can

be achieved and the assumptions of a nondcgenerate two-level

atomic system met, it should offer a very useful diagnostic

technique for the study of combustion.


Analytically, the cw dye laser fluorescence spectrometry

system is considerably different than the low duty cycle

pulsed laser systems used previously by other workers (4,8,30,

31). Practically, it is a less complicated system to work with

because the laser intensity is easily and accurately measurable,

and the beam is convenient to align and focus. Because it is

a cw source, it may be conveniently chopped, and conventional

lock-in detection may be used. At present, the cw dye laser

has the decided disadvantage of very limited wavelength

tunability. Because the majority of analytically useful atomic

fluorescence transitions occur in the blue to ultraviolet

region, this limited tunability in the visible region (approx-

imately 520 700 nm) is a distinct restriction. Nevertheless,

there are a number of useful atomic transitions present in the

visible region that have not been.previously applied for ana-

lytical purposes which may be studied to determine the suitabi-

lity of the cw dye laser as an atomic fluorescence source. A

recent report (32) of a detection limit of 100 atoms cm-3 for

sodium in a quartz vapor cell indicates the potential of the

cw dye laser as an excitation source for analytical atomic


Experimental Procedure

The experimental setup is the same as that shown in

Figure 1 with the exception that the single optical lens in

the fluorescence collection path was replaced with a collimating

lens followed by a dove prism (which rotated the optical image

by 900) and another lens to focus the pencil of fluorescence

onto the slit. The use of the dove prism effected an increase

of almost two orders of magnitude in the signal-to-noise ratio

by allowing efficient collection of the entire pencil of fluo-

rescence generated by the focused laser beam passing through

the flame. In this case, the optimum slit width was a com-

promise between allowing the full width of this pencil to enter

the spectrometer and maintaining a sufficiently narrow spec--

tral bandwidth to minimize flame background noise. For the

analytical studies, a red-sensitive photomultiplier tube

(Hamamatsu R818) was substituted for the 1P28A photomultiplier

tube used previously. As before, Table 1 lists pertinent in-

strumental details.

Standard stock solutions were prepared from reagent

grade chemicals according to the tables given by Smith and

Parsons (33). Necessary dilutions were prepared as required

at the time measurements were being taken. In all cases, de-

ionized water served as the blank. For resonance transitions,

it was necessary to synchronously scan the dye laser and the

detection monochromator across the atomic line. This was ac-

complished by preparing calibration curves for each laser dye

(wavelength micrometer reading vs. wavelength) and then driving

the wavelength micrometer on the laser with an AC synchronous

motor at 1/8 RPM. This provided a scan rate of about 0.01
nm s The stepping motor drive on the detection spectrom-

eter was then driven at the same rate to achieve synchronous

scanning. Because the birefringent tuning element in the dye

laser cavity was not perfectly linear with wavelength, the

scan rate varied from 0.009 to 0.012nm s-1 depending upon the

spectral region of the dye being used. The detection mono-

chromator was adjusted to the proper scan rate by driving its

stepping motor externally with a precision signal generator

of the proper frequency. In general, it was not possible to

scan synchronously over a wavelength range of more than 5 nm

unless great care was used in matching the rate of the detec-

tion monochromator with the rate of the dye laser. Calibration

graphs for rhodamine 6G and sodium fluorescein are shown in

Figure 8.

For nonresonance transitions, the detection spectrom-

eter was tuned to the fluorescence wavelength of interest by

hand and peaked on the background atomic emission signal from

the flame. Then, with the detection spectrometer fixed on

wavelength, the laser was scanned across the region containing

the excitation wavelength while recording the lock-in signal

originating from the fluorescence wavelength.

Operation of the cw dye laser

Figure 9 shows a schematic representation of the cw dye


.4 /

.co. RGS


530 543 550 530 570 530 590 600 610 S20 330 GI4 650

Figure 8. Calibration Graphs for the cw Dye Laser for Rhodamine
6G (R6G) and Sodium 7luorescein (Na-F)

75 rimm LONG RADIUS BS',:FnrlT NT


\.__ .>,-.\ '^/ M OUTPUT



PFUP i rlfnOR

Figure 9 Model 590 Jet Stream Dye Laser

Optical Schematic

laser used in this study. The design is a three mirror, non-

collinearly pumped, astigmatically compensated cavity. Pump

light from a cw argon ion laser (ca. 4 W, all lines) is

focused onto the dye stream by a pump mirror M4. The dye

stream consists of a high viscosity solvent (ethylene glycol)

with an organic dye solution at a concentration of 10- to 10-3

M. The dye is water cooled to increase its viscosity and

improve jet stream stability. Mirrors M1 and 12 are 5 cm and

7.5 cm focal length high reflectors, respectively. Mirror

M3, the output mirror, is flat. Wavelength tunability is

achieved by a birefringent filter tuning element consisting

of three flat and parallel crystalline quartz plates placed

inside the dye laser cavity at Brewster's angle. The plates

form a birefringent filter which has low loss for linearly

polarized light at a particular wavelength. The plates are

oriented so that the optical axis of each crystal is in the

plane of the face and the axes of all the crystals are parallel

to each other. They function as full-wave plates; as the

plates are rotated about a surface normal, the low loss wave-

length changes because the extraordinary index of refraction

changes. When properly aligned, the dye laser operates at an

efficiency (dye laser output/pump laser input) approaching 0.2.

In this study, two dyes were used: (i) rhodamine 6G was used

for wavelengths'between 570 nm and 650 nm; (ii) sodium fluo-

rescein was used between 535 nm and 580 nm. Sodium fluores-

cein required the addition of COT (cyclooctatetraene) to achieve

lasing action (1 2 ml -i of dye solution). The peak out-

put power for each dye was 1.1 and 0.2 W, respectively.

Choice of Analytical Lines

For excited state fluorescence where the absorDtion

transition originates from some state above the ground level,

the important spectroscopic parameters for use in choosing

the best transition for analysis are the relative atomic popu-

lation in the excited level and the transition probability

(A91) for the fluorescence transition. In choosing the tran-

sitions for use in this study, the primary source of data was

the NBS tables of transition probabilities compiled by Corliss

and Bozman (34). Tables of the relative populations of ex-

cited atomic levels were tabulated according to Parsons et al.

(35). Absorption transitions for each of the seventy elements

included in reference 34 were examined over the wavelength

range available with the dye laser (535 650 nm) and, in

general, any lower level containing 0.1% or more of the ele-

ment's total atom distribution at 2500 K was considered. In

most cases, this required that transitions originating above

about 15,000 cm-1 be rejected. Fluorescence transitions

throughout the visible and ultraviolet regions (200 700 nm)

were then examined and those originating from the same upper

levels terminated by the suitable absorption transitions,

with transition probabilities larger than about 10 s were

listed for possible experimental consideration. Table 7

lists the elements ultimately studied along with the pertinent

spectroscopic data. In some cases (most notably barium), it

was possible to choose among several possible combinations of

Spectroscopic Data for the Transition Examined for Analytical Use

Lower Levela


Upper Level a


gA x 108a


% Fopulationb
in Lower Level
at 2500 K






Table 7 continued

Lower Levela


Upper Levela


gA x 108


% Populationb
in Lower Level
at 2500 K




Data taken from reference 34
Data taken from reference 35
DData taken from reference 35


absorption transition with fluorescence transition. In the

case of barium,excitation at any of the first six wavelengths

given in Table 7 resulted in fluorescence at the other five.

The close spacing between the upper levels (AE Z 300 cm-I)

clearly results in very efficient interlevel mixing.

Analytical Results

Limits of detection (LOD) were obtained for eleven ele-

ments by atomic fluorescence with the cw dye laser. In most

cases, the determination of the LOD was performed as the last

step in the measurement of the analytical calibration curve.

Signal-to-noise ratios were obtained at the lowest easily

measured concentrations (typically about an order of magni-

tude above the limit of detection) as the average of four to

six measurements of an analyte solution against a deionized

water blank. The signal-to-noise ratio at this concentration

was then extrapolated to a signal-to-noise ratio of two at

the limit of detection. Table 8 lists the LOD's, flames, and

the excitation and fluorescence lines used for the eleven

elements. Also listed are the best reported literature values

of the LOD for these elements obtained by atomic fluorescence

and atomic absorption flame spectrometry (AAFS) (26). Where

it was practical, analytical calibration curves were obtained

at solution concentrations up to 1000 ug ml-1. Figure 10

shows these plots for barium, sodium, vanadium, neodymium, and

lithium. The growth curves are linear with slopes close to

Table 8
Analytical Results and Spectroscopic Data for the Elements
for Analytical Use

Element 1xcte fl Flamea L.O.D.b L.O.D.C L.O.D.
This Work Conventional AAFS
nm nm Ug ml- pg ml- g ml-

Barium 611.1 606.3 N20/A 0.5 0.22
Barium 553.7 553.7 N20/A 0.04 0.22
Copper 578.2 327.4 N20/A 100 0.005 0.001
Lithium 610.3 610.3 Air/A 0.5 -0.0003
Molybdenum 585.8 345.6 N20/A 14 0.3 0.03
Neodymiun 562.0 562.0 N20/A 2 2 0.6
Rhodium 598.3 339.7 N20/A 0.3 0.15 0.004
Scandium 570.8 568.7 N20/A 50 0.01 0.02
Sodium 589.5 589.5 Air/A 0.0001 0.001 0.001
Uranium 591.5 591.5 N2O/A 500 12
Vanadium 609.0 609.0 N2O/A 0.3 0.07 0.02
Strontium 553.5 550.4 Air/A 0.1 0.01 0.002

aN2C/A = nitrous oxide acetylene flame; Air/A = air acetylene flame.

L.O.D. defined for signal-to-noise ratio = 2.

CFrom reference 26


F 1

.01 II ca3 Or 0C, 0
Analyte Concentration (vg ml-1
Figure 10. Analytical Curves for Sodium, Vanadium, Neodynium, Barium and Lithium

unity. In the case of sodium, the linear.dynamic range of

the analytical curve was observed to be extended at high

laser irradiance. At low laser irradiance unfocussedd) the

sodium analytical curve showed negative curvature at sodium

solution concentrations above 500 pg ml-1

Referring to Table 8, the LOD's are found to be better

than previous AAFS or conventional AFFS results for sodium

and barium where the best resonance line was within the range

of the dye laser. For the other elements, the LOD's are

inferior to previous AAFS or conventional AFFS results although

in some cases (Nd, Rh, V) only slightly so. This is not sur-

prising when one considers the disadvantage of using excited

state levels with low populations and the generally poorer

transition probabilities of the lines used.

For most of the elements, the nitrous oxide acetylene

flame was necessary to achieve both sufficient excited level

populations and greater atomization. For all cases using

this flame, the limiting source of noise proved to be flame

instability related either to scatter of the laser beam from

random particle presence in the flame (in the case of resonance

transitions) or to fluctuations in the flame background con-

tinutum or band emission (in the case of nonresonance transi-

tions). In order to achieve reproducibility, it was necessary

to clean the capillary burner head frequently and to adjust

carefully the gas flow system for optimum signal-to-noise



If the cw dye laser were tunable to the best atomic line

for an element, it should be capable of producing limits of.

detection at least one or two orders of magnitude better than

conventional atomic fluorescence sources. This is due primarily

to the increase in available spectral irradiance achieved

with the dye laser. The usefulness of the cw dye laser is

enhanced by the ability to select unusual atomic or molecular

transitions for analytical use. It appears that unless much

larger argon ion laser pumps are used, the available power in

the blue region of the spectrum will be too low to achieve

full saturation for most elements. However, it is important

to emphasize that once saturation is achieved it is disad-

vantageous to further increase the source intensity since the

fluorescence will remain constant while scatter will continue

to increase. At present, then, the analytical utility of

the cw dye laser is limited primarily by its lack of tunability

throughout the ultraviolet and blue spectral regions where the

more useful atomic fluorescence transitions occur.


The tunable cw dye laser has been shown to be capable

of producing saturation conditions for sodium and to be a

suitable excitation source for analytical atomic fluorescence

flame spectrometry. If the tunable wavelength region of the

cw dye laser is extended, either by frequency doubling or by

use of new dyes or shorter wavelength pump lasers, it should

have considerable usefulness as a versatile atomic fluores-

cence excitation source. At the present level of technology,

however, there is still much fluorescence research for which

this laser may be advantageously used. When saturation is

attained, a measurement of the saturation spectral irradiance,

Es offers a means of determining the quantum efficiency

(equation 30). Because saturation can be achieved for sodium

it should be possible to make accurate measurements of the

sodium quantum efficiency under a variety of flame conditions.

Analytically, the discrete spectral nature and tunability of

the cw dye laser makes it an ideal source for nondispersive

atomic fluorescence. An experimental system using a filter

for wavelength selection, a photomultiplier in close proximity

to the flame, multipass optics and beam expansion should show

considerable improvement in limits of detection for the ele-

ments studied in this work. The cw dye laser should also

prove valuable when used with a nonflame atomization cell,

such as a carbon rod, because of the reduced background and

the ease with which it can be focused and aligned. The

use of intercavity etalons for the selection of a single

mode of the dye laser output offers the possibility of in-

vestigating the effects of a high intensity narrow line source

on atomic fluorescence. As a source for molecular fluores-

cence, the cw dye laser should prove useful for examining

the fine structure of several species which may be present

in flames, in Darticular SrOH and BaO.


1. J. D. Winefordner and T. J. Vickers, Anal. Chem., 36,
161 (1964).

2. C. V. Shank, Rev. Mod. Phys., 47, 649 (1975).

3. J. P. Webb, Anal. Chem., 44, 30A (1972).

4. L. M. Fraser and J. D. Winefordner, Anal. Chem., 43,
1693 (1971).

5. M. B. Denton and H. V. Malmstadt, Appl. Phys. Lett., 18,
489 (1971).

6. L. M. Fraser and J. D. Winefordner, Anal. Chem., 44, 1444

7. N. Omenetto, N. N. Hatch, L. M. Fraser and J. D. Wine-
fordner, Spectrochim. Acta, 28B, 65 (1973).

8. N. Omenetto, N. N. Hatch, L. M. Fraser and J. D. Wine-
fordner, Anal. Chem., 45, 195 (1973)

9. N. Omenetto, G. D. Boutelier, S. J. Weeks, B. W. Smith
and J. D. Winefordner, Anal. Chem., submitted (1977).

10. N. Omenetto, P. Benetti, L. P. Hart, J. D. Winefordner
and C. Th. J. Alkemade, Spectrochim. Acta, 28B, 289 (1973).

11. E. H. Piepmeier, Spectrochim. Acta, 273, 431 (1972).

12. J. Kuhl, S. Neumann and M. Kriese, Z. Naturforsch., 28A,
273 (1973).

13. J. D. Winefordner, V. Svoboda and L. Cline, Crit. Rev.
Anal. Chem., 1, 232 (1970).

14. N. Omenetto and J. D. Winefordner, "Atomic Fluorescence
Spectroscopy with Laser Excitation," Chapter in Analytical
Laser Spectroscopy, N. Omenetto, ed., Wiley-Interscience,
New York, NY, in preparation.

15. D. K. Killinger, C. C. Wang and M. Hanabusa, Phys. Rev. A,
13, 2146 (1976).

16. R. H. Pantell and H. E. Puthoff, Fundamentals of Quantum
Electronics, Wiley, New York, NY (1969).

17. A. P. Thorn, Spectrophysics, Chapman and Hall, London

18. M. Hercher, Appl. Opt., 6, 947 (1967).

19. J. D. Winefordner, S. G. Schulman and T. C. O'Haver,
Luminescence Spectrometry in Analytical Chemistry, Wiley-
Interscience, New York, NY (1972)

20. A. C. G. Mitchel and M. W. Zemansky, Resonance Radiation
and Excited Atoms, Cambridge Univ. Press, Cambridge, MA
(1T96 -.-

21. II. P. Hoomayers, Ph. D. Thesis, Utrecht (1966).

22. W. L. Wiese, M. W. Smith and B. M. Miles, Atomic Transi-
tion Probabilities, Vol. II, Sodium through Catcium, U. S.
Government Printing Office (169).

23. J. J. Barrett and N. I. Adams III, J. Opt. Soc. Am., 58,
311 (1968).

24. H. Haraguchi and J. D. Winefordner, Appl._ pec., sub-
mitted (1977).

25. J. D. Winefordner, M. L. Parsons, J. M. Mansfield and W.
J. McCarthy, Anal. Chem., 39, 436 (1967).

26. M. L. Parsons, B. W. Smith and G. E. Bentley, Handbook of
Flame Spectroscopy, Plenum Press, New York, NY (1975).

27. J. D. Winefordner and T. J. Vickers, Anal. Chem., 36,
1947 (1964).

28. P. J. Th. Zeegers and W. P. Townsend, Spectrochim. Acta,
24B, 243 (1969).

29. P. Lijnse, Ph. D. Thesis, Utrecht (1973).

30. H. L. Brod and E. S. Yeung, Anal. Chem., 48, 344 (1976).

31. S. Neumann and M. Kriese, Spectrochim. Acta, 29B, 127

32. W. M. Fairbank, T. W. Hansch and A. L. Schawlow, J. Opt.
Soc. Am., 65, 199 (1975).

33. B. W. Smith and M. L. Parsons, J. Chem. Ed., 50, 679 (1973).

34. C. 1. Corliss and W. R. Bozman, Experimental Transition
Probabilities for Spectral Lines oi Seventy Elements,
NBS Monograph 53, U. S. Government Printing Offlce (1962).

35. M. L. Parsons, B. W. Smith and P. M. McElfresh, Appl.
Spec., 27, 471 (1973).


Benjamin Willard Smith was born in Safford, Arizona, on

May 10, 1951. He was graduated in June, 1969, from Safford

High School in Safford, Arizona. In December, 1972, he re-

ceived the Bachelor of Science Degree in Chemistry from

Arizona State University in Tempe, Arizona. In January, 1972,

he entered the University of Florida as a Graduate Student

in the Department of Chemistry, College of Arts and Sciences.

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

4ames D. Win fordner, Chairman
Graduate Research Professor
of Chemistry

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

Roge7jG. Bates
Professor of Chemistry

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy

erhard M. Schmicd
Associate Professor of Chemistry


I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

Paul Urone
Professor of Environmental

I certify that I have read this study and that in my
opinion it conforms to acceptable standards of scholarly
presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.

Kuang-Pang Li
Assistant Professor of Chemistry

This dissertation was submitted to the Graduate Faculty of
the Department of Chemistry in the College of Arts and Sciences
and to the Graduate Council, and was accepted as partial
fulfillment of the requirements for the degree of Doctor of

March 1977

Dean, Graduate School

aM46. 77 14 6

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs