EPA-660/2-73-007
August 1973
Environmental Protection Technology Series
Evaluation Of Flame Emission
Determination Of Phosphorus In Water
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five broad
categories were established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
•3. Ecological Research
<*. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
degradation from point and non-point sources of
pollution. This work provides the new or improved
technology required for the control and treatment
of pollution sources to meet environmental quality
standards.
-------
EPA-660/2-73-007
August, 1973
EVALUATION OF FLAME EMISSION
DETERMINATION OF PHOSPHORUS IN WATER
by
William Rudolf Seitz
Southeast Environmental Research Laboratory
College Station Road
Athens, Georgia 30601
R.OAP 16ADN-31
Program Element 1B1027
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97330
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20102 - Prlca 60 cents
-------
ABSTRACT
i
NUCOR's flame spectrometer for phosphorus analysis was
evaluated. Response to phosphorus in the form of H3PO4
was linear from 3 pg/liter, the detection limit, to 120
mg/liter, the highest concentration tested. .Metal ions
depress phosphorus emission and must be removed by
cation exchange prior to analysis. High concentrations
(£5 mg/liter) of sulfur interfere positively. Volatile
phosphorus compounds produce a larger signal for a given
phosphorus concentration than nonvolatile compounds.
River water samples were spiked with inorganic and
organic phosphorus and analyzed. The measured phos-
phorus concentrations were 10-25% lower in river water
than in deionized water.
ii
-------
CONTENTS
Page
Abstract ii
Acknowledgment iv
Sections
I Conclusions 1
II Recommendations 2
III Introduction 3
IV Experimental 4
V Results and Discussion 7
VI References 17
111
-------
ACKNOWLEDGMENTS
George Yager should receive credit for expert plumbing
assistance, running the quality control samples/ and
correctly diagnosing and eliminating the cause of
frequent flameouts.
IV
-------
SECTION I
CONCLUSIONS
Flame emission spectrometry is suitable for determining
dissolved phosphorus in natural waters if metal ions are
first removed by treatment with a cation-exchange resin.
Total phosphorus analysis requires a method for
solubilizing particulate phosphorus.
Sulfur is a possible interference in phosphorus
determinations.
The spectrometer delivered to the EPA is hot completely
satisfactory for routine operation, because background
and sensitivity drift with time.
-------
SECTION II
RECOMMENDATIONS
The flame spectrometer should be modified to minimize
drift in background and sensitivity.
A simple chemical procedure for solubilizing particulate
phosphorus must be developed to make total phosphorus
analysis possible.
-------
SECTION
INTRODUCTION
The total phosphorus concentration in a body of water is
a measure important to studies of eutrophication (1).
The standard method for total phosphorus is to convert
all phosphorus to orthophosphate by wet acid digestion,
followed by colorimetric determination of orthophosphate
using the reaction with ammonium molybdate to form
molybdophosphoric acid (2) .
Flame emission spectrometry is a possible alternative
method for total phosphorus that would not require
conversion of all phosphorus to orthophosphate. In a
cool hydrogen-air flame, phosphorus forms POH which
emits a broad band spectrum peaking at 525 nm. (3).
This emission is used to selectively detect phosphorus-
containing gas chromatography effluents (4) and can
also be used to measure phosphorus in detergents (5,6),
lubricating oil (7) and rocks (8). However, the only
reported method with sufficient sensitivity for phos-
phorus analysis in water requires a nebulization chamber
heated to 90QO C (9) .
A flame emission spectrometer, built specifically for
total phosphorus analysis in water by NUCOR Corporation
under EPA contract, achieves the additional sensitivity
needed for water analysis by ultrasonic nebulization to
convert the sample to an aerosol, rather than pneumatic
nebulization as in conventional flame spectrometry.
With ion exchange pretreatment to remove interfering
cations, the spectrometer successfully measures dissolved
phosphorus at concentrations likely to be found in nature,
It does not respond to particulate phosphate.
-------
SECTION IV
EXPERIMENTAL
SPECTROMETER
The flame emission spectrometer is described in detail
in the final report for the contract between NUCOR
Corporation and the EPA CIO). The essential components
of the spectrometer are an air-cooled burner, an inter-
ference filter to selectively pass the emission at 525
nmf and a photomultiplier to measure emission intensity.
The burner is separated into a flame chamber and an
emission chamber by a 1 mm orifice. This confines the
flame to an area not seen by the photomultiplier, thus
reducing the flame's contribution to background light
intensity.
Hydrogen is supplied directly to the burner. Air is
drawn into the burner by using an air driven aspirator
to create an area of reduced pressure downstream from
the burner. Because both the hydrogen flow and the air
flow through the burner are drawn into the area of
reduced pressure, the flow of air to the burner varies
inversely with hydrogen flow for a given aspirator
setting.
The gas flow system on the spectrometer delivered by
NUCOR was modified for the evaluation. House air was
used to drive the aspirator and cool the burner because
the compressor supplied by NUCOR was too noisy. The
compressor could be used for field work. Hydrogen,
aspirator air, and cooling air were all monitored by
flowmeters controlled by needle valves. No attempt was
made to measure air flow to the burner because inserting
a flowmeter in the air line would reduce the flow rate.
The flame was lit by reducing the air flow to the
aspirator, turning on a spark generator, and gradually
increasing the hydrogen flow. Lighting of the flame is
indicated by a characteristic "popping" sound accom-
panied by an increase in background signal. Immediately
after the flame ignites, the air flow through the
.aspirator is increased to the normal operating level.
Throughout most of the evaluation, the flame went out
spontaneously every 30-45 minutes except when the
burner was unusually hot. This happened when enough
-------
water vapor condensed in the burner to block the 0.31 cm
i.d., 5 cm tube from the emission chamber to the
aspirator. The problem was later corrected by enlarging
the tube diameter.
NEBULIZER
Sample aerosol is generated ultrasonically by a Mistogen
EN142 electronic nebulizer normally used for inhalation
therapy. The aerosol is blown out through a flexible
plastic tube reduced to 0.62 cm i.d. at the outlet. The
outlet is aimed at the sample inlet port of the spec-
trometer and positioned 4 cm away. It is clamped onto
a ringstand held rigidly in position. When the
aspirator air is turned on in the spectrometer, aerosol
is drawn into the burner.
The rate of aerosol generation varies from sample to
sample (11). However, this does not affect observed
emission intensity as long as the rate at which aerosol
comes from the nebulizer exceeds the rate at which
aerosol is drawn into the burner.
Sample can be added directly to the nebulizer (requiring
100-150 ml for aerosol generation), or it can be added
to medication cups that fit into the nebulizer vessel,
thus reducing the minimum sample requirement to 10 ml.
Both the nebulizer vessel and the medication cups are
plastic, which can lead to problems with phosphorus
compounds that tend to adsorb on plastic.
PROCEDURES
Interference studies and calibrations of emission vs.
phosphorus concentration were done by making standard
additions to 250 ml of solution in the nebulizer
vessel. The nebulizer vessel was calibrated by weighing
out 250 grams of water. It was then used volumetrically.
Response to phosphorus in different compounds was
measured and analytical quality control samples were
analyzed using the medication cups.
Because 10-20 seconds are required for the aerosol to
stabilize, the first 30 seconds of aerosol flow were
blocked from the spectrometer by a paper towel. This
led to sharper emission peaks and reduced the amount of
-------
water collecting in the sample inlet through aerosol
condensation. When drops form in the sample inlet,
they must be removed by a paper towel so that the
quantity of aerosol reaching the flame is not reduced.
Sample cups can be exchanged readily and total time for
one emission reading is approximately 60 seconds.
Phosphorus standards were prepared by weighing. Dilute
standards, <25 ing/liter, were prepared fresh daily.
-6
-------
SECTION V
RESULTS AND DISCUSSION
FLAME CONDITIONS
Figure 1 shows phosphorus emission intensity and
background intensity as a function of air flow through
the aspirator for fixed hydrogen flows. At low flow
rates increasing aspirator air flow increases the
observed phosphorus emission, probably by drawing more
aerosol into the flame. Above a certain flow rate,
increasing the aspirator air leads to a decrease in
phosphorus emission and a sharp increase in background.
The decrease in phosphorus emission occurs because the
flame is no longer hydrogen rich, the condition favoring
POH emission (12, 13). Because air flow to the burner
varies inversely with hydrogen flow, the optimum aspira-
tor flow rate is greater for higher hydrogen flow rates.
Normal operating conditions were hydrogen flow 800-1000
cc/min., aspirator air 8-9 liters/minute and cooling air
1 liter/minute.
RESPONSE TO PHOSPHORUS
Emission intensity vs. phosphorus concentration is linear
from 0.003 mg/liter, the detection limit, to 120 mg/liter,
the highest concentration tested, for phosphorus in the
form of H3P04. Figure 2 shows raw data over the range
0-0.15 mg/liter.
The blank emission for deionized water is usually less
than the background emission in the absence of aerosol.
Increasing the hydrogen/air ratio increases the water
blank relative to background.
Response to phosphorus is most sensitive immediately
after the flame is first lighted, i.e. when the spec-
trometer is coolest. As the instrument warms up, the
sensitivity for phosphorus decreases. At least 15
minutes warmup is necessary to achieve sufficient
stability to perform an analysis.
Prolonged continuous operation of the spectrometer (over
one hour) leads not only to a loss of sensitivity but
-------
a)
LUtT)
^P"^
Q
a:
o
^:
o
UJ
(7)
Q:
6 8 10
ASPIRATOR AIRt liter/min
ugh aspirator
ission for
o e
rH CD
,f*
•P 13
£ §
0 0
rH >-|
MH O>
J<^
rH U
-H tO CO
ttf tQ IS
.
£
•H
^
0
0
O
•«*
in
•
C
•H
,6
0
O
O
O">
p-
.
c
•H
e
o
o
0
r~>
0
•H
J
£
O
•rH
CO
CO
. -H
c
0
•rH
to
CO
•H
C C «H.
•H (0
>» c
CO 0)
3 tyi
> ^i O
O rH
M-i X! 'd
Q P! ^
CO X!
JJ O
O x! T3
-------
I—Zmin.H
Figure 2. Raw data showing phosphorus emission as a
function of concentration from 0 to 0.15
ing/liter, in' 0.03 ing/liter increments
-------
also to a substantial increase in background. The
background remains high when the flame is turned off.,
indicating that heating of the photomultiplier tube is
responsible. Modification of the spectrometer design
to eliminate this problem should be simple.
Table 1 lists relative emission per unit phosphorus
for several phosphorus containing compounds. Volatile
phosphorus compounds give a larger signal than non-
volatile compounds. This is consistent with'the
observation that response to vapor phase triethyl phos-
phate is over 10 times more sensitive than the response
to aqueous triethyl phosphate for the same amount of
phosphorus reaching the flame per unit time CIO). As
aerosol evaporates in the flame, the volatile phosphorus
compounds presumably vaporize sooner than the non-
volatile compounds leading to more efficient excitation
of POH emission.
A series of phosphorus-containing pesticides were to be
run, but these compounds adsorbed on the plastic medica-
tion cups so rapidly that their phosphorus emission was
impossible to measure satisfactorily. Of the compounds
tried, only ruelene was stable. Triethyl thiophosphate
adsorbed fairly rapidly, so the emission reading was
taken immediately after the dilute solution was prepared,
The medication cups used with adsorbing organic phos-
phorus compounds slowly desorbed phosphorus during sub-
sequent use and had to be thrown out.
Because the organic compounds dissolved in natural
waters are not likely to be volatile (14-16), the fact
that volatile compounds give greater response per unit
phosphorus is probably not a problem for phosphorus
analysis in natural waters.
SULFUR INTERFERENCE
The hydrogen-air flame efficiently excites sulfur 82
molecular emission peaking at 394 nm. (3,4,9,12,13).
Some sulfur emission occurs at 525 nm, the wavelength
of maximum phosphorus emission. Because two atoms
produce one emitting 82 molecule, emission intensity
varies with the square of the sulfur concentration.
Figure 3 plots the relative emission intensities for
phosphorus and sulfur as a function of sulfur concentra-
tion. Above 5 mg/liter, sulfur rapidly becomes a
significant interference.
10
-------
Table 1. EMISSION INTENSITY PER UNIT PHOSPHORUS FOR
DIFFERENT COMPOUNDS (relative to H3P04)
Compound
H3P04
Na4P207
Sodium
glucose
phosphate
Ruelene*
Triethyl
phosphate
Triethyl
thiophosphate
Emission
Intensity
1.0
1.0
1,0
1.3
1.7
3.2
Volatility
b.p. 215°
95°/10 mm
b.p. 1000/
16 nun
*The structure of ruelene is
CH3-0-P-N,
11
-------
1000
ID
a:
O
i
a.
C/)
O
I
Q_
o:
o
UL
O
LJ
_l
LL)
CO
CD
O
100
10
LOG SULFUR CONCENTRATION, mg/liter
100
(0
tf
a
-H
Miii
^^i
^ 3
6 0-M
0) -H -P
•P O
(DUO)
6 C rH
14-1
O CO C
(0 O
>i -H
4J W -P
•H ^ (0
^^ ii t§
•H O -P
•P A C
o a CD
0) W O
rH O G
a) x: o
w a. o
n
cu
g.
•H
CO
O
.C
a
CO
W
§
en
CO
QJ
(U
-------
If sulfur emission is suspected, its presence can be
detected by diluting the sample to determine whether
response is first or second order.
CATION INTERFERENCES
Cation depression of POH emission has been reported
previously (5-9, 13). Figure 4 shows the effect of
added Ca and Mg on phosphorus emission. Even Na and K
depress phosphorus emission.
Interfering cations were removed by exposing a sample to
20-50 mesh Amberlite IR-120 strongly acidic cation
exchange resin. Resin performance was tested using a
column containing approximately five grams of resin. It
was verified that passing deionized water through the
column did not change the blank. :One mg/1'solutions ot
phosphorus in the form of H3PO4 and sodium glucose phos-
phate went through the column without any change in
observed emission intensity. One riig/1 phosphorus "as
H3PO4 in the presence of [100 mg/I each" of ~fla,~K, "Ca and
Mg went through the column with less than a"10%" loss in
emission intensity indicating that the column was satis-
factorily removing interfering cations.
RIVER WATER ANALYSIS
To verify that flame emission analysis for phosphorus
worked on river water samples, a series of analyses was
performed on nutrient reference samples provided by
EPA's Analytical Quality Control Laboratory, Cincinnati,
Ohio. These samples are supplied as concentrates to be
diluted by a factor of 200. Each sample type was
diluted both in deionized water and in filtered river
water. Samples 1 and 2 contain inorganic phosphorus and
samples 3 and 4 contain organically bound phosphorus.
Passing river water through the cation exchange column
resulted in losing the spike both for inorganic and
organic phosphorus, even though no loss of phosphorus
was observed in deionized water samples. If the ion
exchange resin was added directly to the medication cup,
it depressed the rate of aerosol generation. The pro-
cedure adopted was to add ion exchange resin to a liter
of river water and shake it. This water was used for
the river blank and the sample. The spiked river water
was again treated with ion exchange resin to remove
13
-------
3
J3
cc
o
o
X
CL
O = Co
A = Mg
w
§
•H
•H
W
0)
«J
e
•H o
o -H
i-l W
m en
O-H
-o
-------
metal ions accompanying the spike. The spiking was done
just before analysis so that the sample would not sit
more than 10-15 minutes before being run.
Table 2 shows the results. The nominal concentration is
given in the literature accompanying the quality control
samples. The percentage recovery is the.measured con-
centration of the spike in river water over the measured
concentration of the spike in deionized water. Some of
the spiked phosphorus is lost. I believe that it
associates with particulate matter in the river water
and is no longer available for excitation in the flame.
This explanation is consistent with the known behavior
of phosphate in natural waters (17-18). It-may explain
why the spike was lost using an ion exchange column,
which would filter the sample.
From this behavior and the known susceptibility of the
cool hydrogen-air flame to interference, it may be
concluded that the flame does not possess sufficient
thermal energy to break up particulate matter and excite
particulate phosphorus emission.
The measured spike concentrations in deionized water
agree with the .nominal concentrations within 10%.
Sources of variation in measured phosphorus concentration
include:
• • variations in sensitivity with time.
• differences in the quantity and droplet
size of the aerosol with different
solutions.
• uncertainties in peak intensity
measurements.
15
-------
Table 2. ANALYSIS RESULTS
Sample
Number
1
2
3
.4
Form of
Phosphorus
Inorganic
Inorganic
Organic
Organic
Nominal
cone.
(mg/liter)
_*
0.300
0.170
0.85
Measured
Cone. Deion.
H20
(mg/liter)
0.05*
0.28
0.27
0.17
0.19
0.79
0.83
Measured
Cone . River
Water (ppm)
(mg/liter)
0.038, 0.41
0.044
0.22
0.21
0.15
0.17
0.69
0.69
Percent
Recovery
82%
78%
88%
85%
Source of
River Water
Filtered Oconee
River
Filtered Oconee
River
Filtered Oconee
River
Filtered Oconee
River
For samples 2, 3 and 4 the river water was run twice using separate samples. The
same spiked deionized water was used for both measurements. The river water blank
was equivalent to about 0.04 mg/1 phosphorus.
* We ran out of quality control standard before arriving at a suitable analytical
procedure. Data are for a synthetic standard.
-------
SECTION VI
REFERENCES
1. Bartsch, A. F. Role of Phosphorus in
Eutrophication. Environmental Protection Agency/
Corvallis, Ore. Publication Number R3-72-001.
2. Standard Methods for the Examination of Water and
Wastewater. 13th edition. American Public Health
Association, 1971. p. 518-534.
3. B. Draeger, Heinrich Draegerwerk. West Germany
Patent 1,133,918. July 26, 1962.
4. Brody, S. S., and J. E. Chaney. Flame Photometric
Detector-Application of a Specific Detector for
Phosphorus and for Sulfur Compounds. J Gas
Chromatog. £: 42-46, February 1966.
5. Syty, A. Determination of Phosphorus in Detergents
by Flame Emission Spectrometry. Anal. Lett. £:
531-536, August 1971.
6. Elliot, W. N., and R. A. Moslyn. The Determination
of Phosphate in Detergents by Cool-Flame Emission
Spectroscopy. Analyst (London). 96: 452-456,
June 1971.
7. Elliott, W. N., C. Heathcote, and R. A. Moslyn.
Determination of Phosphorus in Lubricating Oils by
Cool-Flame Emission Spectroscopy. Talanta. 19:
359-363, 1972.
8. Syty, A. Determination of Phosphorus in Phosphate
Rock by Cool Flame Emission Spectrometry. At. .
Absorption Newslett. 12: 1-2 January-February 1973.
9. Aldous, K. M., R. M. Dagnall, and T. S. West. The
Flame-Spectroscopic Determination of Sulphur and
Phosphorus in Organic and Aqueous Matrices by Using
a Simple Filter Photometer. Analyst (London). 95;
417-424, May 1970.
10. NUCOR Corporation. Automated Water Monitoring
Instrument for Phosphorus Contents. Final Report
for EPA Contract No. 68-01-0111, 1973.
17
-------
11. Stupar, J., and J. B. Dawson. Theoretical and
Experimental Aspects of the Production of Aerosols
for Use in Atomic Absorption Spectroscopy. Appl.
Opt. 2.' 1351-1358, July 1968.
12. Syty, A. and J. A. Dean. Determination of
Phosphorus and Sulfur in Fuel-Rich Air-Hydrogen
Flames. Appl. Opt. 7^ 1331-1336, July 1968.
13. Dagnall, R. M., K. C. Thompson, and T. S. West.
Molecular-Emission Spectroscopy in Cool Flames.
Analyst (London) . 93_: 72-78, February 1968.
14. Gales, M. E., E. C. Julian, and R. c. Kroner.
Method for Quantitative Determination of Total
Phosphorus in Water. J. Am. Wat. Wks. Assoc. 58:
1363-1368, October 1966. —
15. Fuhs, G. W. Determination of Particulate Phosphorus
by Alkaline Persulfate Digestion. Intern. J.
Environ Anal Chem. I: 123-129, 1971.
16. Burton, J. K. Problems in the Analysis of Phos-
phorus Compounds. Water Research. 7_: 291-307, 1973.
17. Chen, Y. S. R., J. N. Butler, and W. Stumm.
Kinetic Study of Phosphate Reaction with Aluminum
Oxide and Kaolinite. Environ. Sci Technol. 7:
327-332, April 1973.
18. Syers, J. K., R. F. Harris, and p. E. Armstrong.
Phosphate Chemistry in Lake Sediments. J. Enviror
Quality. 2: 1-14, January-March 1973.
18
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
1. Re^artNo.
2.
3. Accession.No.
w
4. Title EVALUATION OF FLAME EMISSION DETERMINATION
OF PHOSPHORUS IN WATER
7. Autbor(s)
Seitz, W. Rudolf
5. htportDate
8. I rforrai.-g Orga:.^etion
Report No.
9. ' Organization
U. S. Environmental Protection Agency
Southeast Environmental Research Laboratory
Athens, Georgia 30601
10. Project No.
16ADN-31
11. Contract/Grant No.
/. Typt i Repc and
Period Covered
12.
nsoric Organ ition
15. Supplementary Notes
Environmental Protection Agency Report Number, EPA-660-2-73-007,
August 1973
16. Abstract
NUCOR's flame spectrometer for phosphorus analysis was evaluated.
Response to phosphorus in the form of H3P04 was linear from 3 yg/liter,
the detection limit/ to 120 mg/literf the highest concentration tested.
Metal ions depress phosphorus emission and must be removed by cation
exchange prior to analysis. High concentrations (>5 mg/liter) of sulfur
interfere positively. Volatile phosphorus compounds produce a larger
signal for a given^phosphorus concentration than nonvolatile compounds.
River water samples were spiked with inorganic and organic phosphorus
and analyzed. The measured phosphorus concentrations were 10-25% lower
in river water than in deionized water.
17a. Descriptors
*Phosphorus, *Phosphates, *Flame Photometry, Phosphorus Compounds,
Organophosphorus Compounds, Inorganic Compounds, Analytical Techniques,
17b. Identifiers
Flame Emission, Emission Spectrometry
17c. CO WRR Field & Group Q5A
18. A vailability
19. Security Class.
i. St
;
rityC
21.
No. of
Pages
Pr "
Send To:
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C. ZO24O
Abstractor W. Rudolf Seitz | institution EPA, Southeast Env. Res. Laboratory
WRSIC IO2 (REV. JUNE 1971)
------- |