Environmental Monitoring Series
Automated Water Monitoring
Instrument for Phosphorus  Contents
                              \

                               Office of Research and Monitoring
                               U.S. Environmental Protection Agency
                               Washington, D.C. 20460

-------
             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
   4.  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 Review Notice

This report has been reviewed by the Office of Research and
Monitoring, EPA, and approved for publication.  Approval does
not signify that the contents necessarily reflect the views
and policies of the Environmental Protection Agency, nor does
mention of trade names or commercial products constitute
endorsement or recommendation for use.

-------
                                          EPA-R4-73-026

                                          June,  1973
     Automated Water Monitoring  Instrument

             For Phosphorus Contents

                        by

                Manfred J.  Prager



             Contract No. 68-01-0111

                Project 16020-GSB

             Program Element 1B1027




                 Project Officer

               Dr. Thomas B. Hoover
  Southeast  Environmental Research  Laboratory
            College  Station Road
            Athens,  Georgia 30601



                  Prepared for

       Office of Research and Monitoring
     UoS. Environmental Protection  Agency
           Washington,  D.  C.  20460
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402
           Price 50 cents domestic postpaid or 35 cents GPO Bookstore

-------
                        ABSTRACT


A prototype instrument was developed by NUCOR Corporation,
Denville, New Jersey, for automatically monitoring total
phosphorus in water.  The analytical principle employed was
flame emission photometry-  Phosphorus compounds burned in
a hydrogen flame emit at about 525 millimicrons.

Conditions were established for the sensitive measurement of
phosphorus in water.  Operating parameters investigated in-
cluded fuel and air flow rates, burner configuration, opera-
ting temperature, method of sample aerosolization, etc.

Using an ultrasonic nebulizer to aerosolize samples of tri-
ethylphosphate in water, it was possible to detect phosphorus
at a concentration of less than 2 parts per billion.  A pro-
cedure was worked out for distinguishing between organic and
inorganic phosphorus with ion exchange resins.  In measurements
designed to determine interference by sodium and calcium, it
was observed that the method is about 1000 times more sensi-
tive towards phosphorus than towards sodium and 5000 times
more sensitive towards phosphorus than towards calcium.

A prototype instrument was designed, fabricated, tested, and
delivered to EPA, Southeast Environmental Research Laboratory.

This report was submitted in fulfillment of Contract No. 68-
01-0111 under the sponsorship of the Environmental Protection
Agency.  Work was completed as of March, 1973.
                          IX.

-------
                         CONTENTS








Section                                              Page




I           Conclusions                                1.




II          Recommendations                            2.




Ill         Introduction                               3.




IV          Experimental                               6.




V           Discussion                                24.




VI          Acknowledgements                          25.




VII         References                                26.
                           111.

-------
                         FIGURES

                                                       Page

1.   Schematic, Hydrogen Flame Emission Water Monitor     4.

2.   Response to Phosphorus Compound Vapors in Air       10.

3.   Response to 0.1 ppm TEP in Water                    15.

4.   Response - Concentration Relationship for TEP
               in Water                                 17-

5.   Block Diagram - Electronic Circuitry                21.

6.   Instrument Wiring Diagram                           22.
                          IV.

-------
                        SECTION I
                       CONCLUSIONS
1.   Flame emission photometry is a suitable method for
    monitoring phosphorus in water.

2.   The method can be adapted to automated instrumentation.

3.   It is possible to detect less than 2 parts per billion
    of phosphorus in water.

4.   By the use of an ion exchange pretreatment, it is possi-
    ble to separately determine inorganic and organic phos-
    phorus .

5.   The method is about 1000 times more sensitive towards
    phosphorus than towards sodium and 5000 times more
    sensitive towards phosphorus than towards calcium.
                            1.

-------
                       SECTION II
                     RECOMMENDATIONS
Additional work is recommended to more fully investigate the
method, to improve the accuracy and reproducibility of the
principle and to improve the design and the reliability of
the instrument.

Specifically the following work is recommended:

1.  Determine response for a variety of phosphorus compounds
    to determine whether response depends only on phosphorus
    content or is affected by compound structure.

2.  Determine possible interference from a variety of ele-
    ments at concentrations encountered in the field.

3.  Determine approximate size and concentration range of
    particulate matter that can be tolerated in the water
    sample.

4.  Provide for direct introduction of a water sample of 10
    ml or less.

5.  Improve response by achieving more reproducible background
    signals or background cancellation.
                            2.

-------
                   III.  INTRODUCTION


The purpose of this project, performed by NUCOR Corporation,
Denville, New Jersey, was to develop and fabricate a proto-
type, automated water monitor for trace quantities of phos-
phorus compounds, based on hydrogen flame emission spectro-
scopy.

Measurement of phosphorus compounds in water is of interest
since pollution of water by such substances can readily occur
from a variety of sources.  These include leaching from ores,
decomposition of organic matter, industrial wastes, deter-
gents, pesticides, and others.  The phosphorus may be present
in inorganic and/or organic form.  Undesirable effects in
water can be attributed to various phosphorus compounds:
many organic phosphorus compounds are toxic and persistent,
polyphosphates interfere with coagulation, phosphates promote
growth of algae, etc.  It is desirable, therefore, to detect,
measure, and identify phosphorus compounds in water in order
to ascertain the need for and to evaluate the effectiveness
of, control measures.

In view of the potential hazards due to the presence of phos-
phorus compounds in the environment, considerable effort has
been devoted to the development of automated monitoring
instrumentation for the measurement of phosphorus compounds,
both in air and water.  One of the most attractive analytical
principles for this application is flame emission spectro-
scopy, because it is highly sensitive and selective and rela-
tively easy to adapt to automated instrumentation.  The method
was used for this work.

This principle, the instrumentation required, and application
to pollution monitoring have been described.    Briefly, when
phosphorus compounds are introduced into a cool, fuel-rich
hydrogen flame, POH radicals are formed.  These produce a
green emission in a region that extends beyond the flame.
The spectral band, with peak emission at about 526 nm, may be
isolated with an interference filter and measured with a pho-
tomultiplier tube.  The presence of phosphorus in a sample can
be detected by monitoring the emission at 526 nm, and the
phosphorus concentration in the sample can be determined by
measuring the intensity of this emission.

One apparatus suitable for such measurements is shown in Fig.
1.  Hydrogen fuel from a cylinder is fed into a burner through
a capillary tube and burns at the tip of this tube.  An inter-
ference filter is mounted next to the burner, opposite the
emission region and the photomultiplier tube is in back of the
filter.   Combustion air and sample are introduced into the
                             3.

-------
/2 OU/fHETSt
                                                                           ftC
                      HYDROGEN  FLAME EMISSION WATER MONITOR




                                     FIGURE 1






                                         4.

-------
burner in the vicinity of the flame.  In the particular
arrangement shown, the burner is attached to the side arm
of an air driven aspirator.  The sample inlet is, there-
fore, at a pressure somewhat below ambient, so that the
combustion air and sample are sucked into the burner.  A
compressor supplies air, part of which is fed to the aspira-
tor and another portion is used to cool the burner.  A high
voltage DC power supply furnishes the voltage required to
operate the photomultiplier tube.  The photomultiplier tube
current is amplified and the signal is displayed on a meter
and may also be recorded.

Continuous, automated flame emission air monitors for phos-
phorus have been under development in this country by various
DOD agencies for more than ten years.  Instruments with dif-
ferent interference filters to permit emission measurements
at about 390 nm are in use to monitor air for sulfur com-
pounds.  Using an instrument developed by the Navy for air
monitoring for phosphorus, some work was earlier done to
investigate the feasibility of detecting phosphorus compounds
in water.   The development of an automated water monitor for
phosphorus compounds which is suitable for field use has not
yet been reported in the literature.

It is the purpose of this work to develop such an instrument.
Specifically, it was desired to:

1.  Develop a burner capable of operating effectively during
    exposure to water samples.

2.  Develop a sample introduction system to achieve high
    sensitivity and reliable burner operation.

3.  Develop the necessary optical and electronic components
    and fabricate a working prototype instrument.

4.  Evaluate the instrument in the laboratory and determine:

    a.  Precision and limits of phosphorus measurement.

    b.  Discrimination between organic and inorganic phos-
        phorus compounds.

    c.  Determine possible interferences, especially sodium
        and calcium and the feasibility of using the instru-
        ment for measuring phosphorus in sea or estuarine
        water.
                            5.

-------
                     IV.  EXPERIMENTAL
Laboratory Study Instruments

The laboratory investigations prior to prototype fabrication,
were performed with two instruments based on the schematic
shown in Fig. 1.  One was a stainless steel housing in which
were mounted the burner, aspirator, interference filter, and
photomultiplier tube.  A second instrument employed an alumi-
num cabinent which contained the same components plus ignitor
circuit, air control valve, plus additional electrical con-
nectors and pneumatic fittings and lines.  Because of diffi-
culties in properly aligning the burner and air inlet fittings
in this second instrument, the former one was used in most of
the studies.

The burner, aspirator, and housings were designed and fabri-
cated at NUCOR.  Major commercial components included:

Interference filter:  Baird - Atomic Inc., Catalog #11-92-2
with peak wavelength at 5250 Angstroms, transmittance at
peak wavelength 50-70 percent and bandwidth at half peak
transmittance 150 Angstroms.

Photomultiplier tube:  RCA #1P21.

Photomultiplier tube power supply:  John Fluke Mfg. Co.,
Model 412A variable high voltage DC supply.

Electrometer:  Keithley Instruments, Inc., Model 610C.

Compressor:  Cast Mfg. Corp., Model 1531-107-G288.

Recorder:  Hewlett Packard, Model 680 5-in. strip chart
recorder.

Ignitor circuit:  Ignition was achieved with a high voltage
spark using a circuit obtained from Ridge Electronics Corp.
West Millington, N.J.

Compressed hydrogen gas was used as the fuel.  Ambient air
was used to support combustion and the sample was fed into
this air stream and carried by it to the burner.  Air was
sucked into the burner because of a slight negative pressure
at the air inlet created by the air driven aspirator.  The
aspirator was heated with a cartridge heater to prevent
water condensation.

The metal burner consisted of two chambers - the flame cham-
ber and the emission chamber.  A 1/4 in. o.d. brass tube was
                             6.

-------
screwed into the flame chamber and served as the air and
sample inlet.  A piece of teflon rod with a hole drilled
into it to accommodate a length of 17 gauge hypodermic
needle tubing was also inserted into the flame chamber.
The hydrogen was introduced into the burner through the
hypodermic needle and burned at the end of the capillary.
Flame ignition was achieved by sparking between the hypo-
dermic tubing and the burner housing.

Since the emission is formed beyond the flame, the flame
itself was not viewed in order to reduce the background
signal.  Coloration of the flame by extraneous substances
as well as flame flicker would contribute to the background
and lower the signal to noise ratio if viewed by the photo-
multiplier tube.  To eliminate the flame from view, the flame
and emission chambers were separated by 0.040 inch orifice
through which the gases, but not the flame, could penetrate.
A Pyrex glass tube was fitted into the emission chamber
housing.  The green phosphorus emission was visible in this
chamber and was viewed by the photomultiplier tube through
the interference filter.  The emission chamber was connected
to the aspirator through which the gases were exhausted.

A portion of the air from the compressor was passed over the
exterior of the burner in order to cool the burner. If this
is not done, phosphorus sensitivity is decreased.

Generation of Phosphorus Containing Vapors.

Initial evaluation of burner, optics, electronic and pneu-
matic components was performed with samples of phosphorus
compounds in air rather than in water.  This permitted
measurement of phosphorus emission signals uncomplicated by
background noise and fluctuations due to the presence of,
and cooling by, water.  On the basis of available vapor
pressure data, triethylphosphate (TEP) appeared suitable
for the generation of useful concentrations of phosphorus
in air.  By appropriate control of temperature and flow
rates, it was possible to obtain with TEP convenient phos-
phorus concentrations that ranged from 10 ppm to 1.0 ppb
in air.

Using TEP vapor pressure data from the literature, vapor
pressure-temperature relationships and the ideal gas equa-
tion one calculates the following phosphorus concentrations
in air, expressed in micrograms per liter of the saturated
vapor:
                             7.

-------
                °C                ug/1

                22                508

                24                583

                25                624

                26                666

                28                761

By air dilution of the saturated vapor desired phosphorus
concentrations in air were generated.

A length of filter paper was loosely supported in a 10 cm
length of 6 mm Pyrex tubing.  The paper was then saturated
with triethylphosphate.  One end of this tube was connected
to a compressed air cylinder to provide a regulated flow of
saturated TEP vapor.  To the other end of the Pyrex tube
was connected a hypodermic needle.  The tip of the needle
was inserted into the instrument air inlet.  From measure-
ments of the ambient temperature, saturated vapor flow rate
and diluent air flow rate at the instrument inlet the phos-
phorus- concentrations fed to the burner could be calculated.
Triethylphosphate was added occasionally to the filter paper
to assure that the air from the cylinder was saturated with
TEP vapor prior to dilution at the instrument air inlet.

Initial Instrument Evaluation with Air Samples.

To evaluate burner performance, triethylphosphate vapors
were added to the combustion air to give a phosphorus con-
centration of about 9 ppm.  The effect of hydrogen and air
flow rates on phosphorus response and background current
were measured.

The data obtained were not sufficiently reproducible so that
individual emission values could be assigned to each set of
experimental conditions, nevertheless, the trend of the data
allowed useful conclusions to be drawn.  At a constant air
flow of 200 ml/min a hydrogen flow of at least 200 ml/min
was required for high sensitivity.  Background response was
independent of hydrogen flow.

-------
Hydrogen     Number
  Flow         of
 (ml/min)  Measurements     Background  (amps)  Signal  (amps)

  150          4            1.0-2.3xlO~8      2.5-4.7xlO~6

  170          3            0.7-3.6xlO~8      1.0-2.1xlO~6

  200          8            0.8-2.4xlO~8      1.3-10xlO~6

  275          3            1.6-3.8xlO~8      4.8-llxlO~
                                      -8               -6
  320          3            1.3-3.5x10        5.7-13x10

  440          3            0.8-2.9xlO~       4.4-12xlO~
At a constant hydrogen flow of 200 ml/min, variation in the
air flow between 150 to 300 ml/min showed the signal to be
independent of air flow, but the background to increase with
increased air flow.  Measurements of the temperature of the
external surface of the burner near the flame showed that
the temperature also increased with increasing air flow.  It
is desirable to keep burner temperature low since this will
keep photomultiplier tube temperature low and, therefore,
its sensitivity high.  In order to keep background and noise
levels low and also to conserve fuel gases, hydrogen and air
flow rates were maintained just high enough to achieve good
sensitivity.

Using hydrogen and air flow rates of 200 ml/min, the effect
of phosphorus concentration on response was determined.  The
data for concentrations between 9 and 0.0009 ppm were plotted
on log-log paper and fell on a straight line except for the
point at the highest concentration, as shown in Fig. 2.
These experiments also showed that it is possible to detect
less than 1 ppb of phosphorus in air.  Sensitivity was judged
to be satisfactory and subsequent response measurements were
made with samples of phosphorus compounds in water.

Measurement of Burner Temperature.

Attempts were made to measure the external burner surface
temperature.  There are several reasons for performing such
measurements.

1.  For good performance the burner must be maintained at a
    relatively low temperature.  If the temperature is per-
    mitted to rise too much, sensitivity to phosphorus
    decreases for reasons already mentioned.
                            9.

-------
                                                          Q_
                                                          D.
                                                    O.I    ^
                                                          o
                                                          h-
                                                          UJ
                                                          o
                                                     0.01
                                                    - 0.001
                                               o
                                               o
                                                          o:
                                                          o
                                                          X
                                                          Q.
                                                          O)
                                                          O
                                                          X
                                                          0-
I.OxlO
    -6
                    .-7
     I.OxlO"'            I.OxlO"

RESPONSE ( AMPS )
.0x10
                                                     0.0001

                                                     -9
                   FIGURE  2


    RESPONSE TO PHOSPHORUS  COMPOUND

           VAPORS  IN AIR
                      10.

-------
2.  For reasons of safety, it is desirable to know that the
    flame is burning and that the fuel is being combusted
    and not escaping into the surrounding atmosphere and
    presenting an explosion hazard.  One way to determine
    that the flame is on is to measure the burner tempera-
    ture.

3.  A high voltage spark is used to ignite the flame.  The
    ignitor is turned off after ignition since continued
    sparking would be detected by the photomultiplier tube
    and interfere with the phosphorus measurement process.
    Also, in case of a flame-out, for reasons of safety, as
    already mentioned, the flame should be relit if the
    hydrogen flow is continued.  It is possible to employ an
    automatic ignition circuit which incorporates a tempera-
    ture sensor and functions when the burner temperature is
    below the operating temperature and shuts off when the
    operating temperature is reached.

4.  During use of the instrument, it is important to know
    that the flame is on so that measurements can be made.
    One way to check on the instrument status is to monitor
    the burner temperature.

Measurements were made of the temperature of the external
surface of the burner.  This was done by clamping a thermo-
couple to the burner with the bi-metallic junction as close
to the flame as possible.  In 12 experiments in which the
burner operated at hydrogen and air flow rates of 200 ml/min
each, the burner temperature ranged from 162 to 180°F.  This
temperature was attained in about 20 minutes after ignition
and could thereafter be maintained within 3°F.  On exting-
uishing the flame, the burner temperature dropped more than
20°F in the first minute.

On the basis of these results, an automatic ignition circuit
was designed.  It consisted of a miniature thermostat that
activated and deactivated the spark ignitor previously de-
scribed and a light indicated ignitor status.  The thermo-
stat which incorporated a bimetal actuated contact was
mounted on the exterior of the burner close to the flame.
The thermostat was in series with the spark ignitor and
light.   The thermostat was closed during ignition and opened
when the burner temperature reached 150°F.  The light and
spark ignitor were then deactivated.  In case of a flame-out
the thermostat closed when the burner temperature dropped to
145°F and reignition occurred.  In practice, it was observed
that the ignitor stopped sparking 7-11 minutes after the
instrument was turned on and also within about 7 minutes
following a flame-out.
                            11.

-------
This thermostat flame-out sensor operated satisfactorily for
about two months until it burned out during an electrical
short.  It was not replaced since flame status could be
monitored simply by noting the .photomultiplier current indi-
cated by the electrometer.

The current observed when the flame was on and there was no
phosphorus in the flame (flame background) was about an order
of magnitude higher than when the flame was out  (dark cur-
rent) .  The flame background was usually l-3xlO~b amps,
while the dark current was usually 2-3x10   amps.  This method
of monitoring flame status has the advantage that a meter in-
dication can be obtained in a few seconds, but the disadvan-
tage that ignition had to be performed manually.  If a flame
out occurred while the instrument was left unattended for
extended periods, appreciable quantities of unburned hydrogen
could escape.  It became evident during the course of the work
that it was possible to design a nebulizer-burner system that
resulted in only very few flame-outs.  It is desirable, how-
ever, to incorporate into a flame emission monitor a flame
out sensor, with an automatic ignitor or alternatively auto-
matic shut-off of hydrogen flow.

Measurement of Response to Phosphorus in Water.

After establishing satisfactory operation of the laboratory
instrument with air samples containing phosphorus, measure-
ments were made of the instrument response to phosphorus in
water.  Triethylphosphate was used for these experiemtns also.
Samples were made up in tap water and also in distilled water
to determine the effect on response of impurities in the
water.  It was anticipated that the sensitivity attainable
would depend largely on the nebulization step.  It was rea-
soned that good performance would require that the sample
must reach the flame in the form of exceedingly fine droplets
that could be vaporized during the short period that they
were in the flame.  It was further thought that the finer
the aerosol particles, the better the chance of reaching the
flame without dropping out, the smaller the effect on flame
temperature, and the less chance of extinguishing the flame.
Liquid samples in flame emission and absorption spectroscopy
are commonly aerosolized with pneumatically driven nebulizers.
Since finer droplets can apparently be obtained with ultra-
sonic nebulizers, it was thought that these may provide
better performance.  Some studies described in the literature
indicate that atomic absorption intensities are greater
usually by factors of 2 to 5, if samples are aerosolized
ultrasonically rather than pneumatically.

It was decided to compare performance obtainable with both
types of nebulizers.  Most of the work employing pneumatic
                             12.

-------
aerolization was performed with the nebulizer used  in  the
Jarrell-Ash flame emission/atomic absorption spectrophoto-
meter.  The ultrasonic nebulizer used was the Mistogen
Equipment Company, Model EN142B electronic nebulizer.  This
unit is used primarily in hospitals and homes to produce
finely atomized medications  for inhalation therapy.

To the exit nozzle of the aerosol breathing tube of the
ultrasonic nebulizer was attached a hose nipple in order to
funnel the aerosol into a narrower stream for impingement on
the instrument's sample inlet.  The nipple was maintained
about 2 inches from the sample inlet.  The pneumatically
nebulized aerosol stream was fed either directly into  the
burner's sample inlet or alternatively was first fed at
right angles to the instrument through an atomizer flask
to remove large droplets and prevent these from entering
into the burner.  Various configurations of atomizer flasks
were used.

The burner was operated initially with water samples under
conditions which had previously resulted in excellent  sen-
sitivity toward phosphorus in air.  However, when water
samples were sprayed into the burner with a pneumatic
atomizer, only those containing as much as one part per
thousand of triethylphosphate could be detected, whether
the spray was aimed directly at the burner or whether  an
atomization flask intervened.  The measured emission from
samples of phosphorus in water  was, therefore, much less
than from air samples containing the same weight of phos-
phorus.  Possible causes for the poor sensitivity were
thought to be:   (1)  too little of the aerosol reached the
flame  (2)  too little of the aerosol was vaporized in  the
flame because  (a)  droplets were too large (b)  residence
time was too short  (c)  flame was too cool.

It was occasionally observed that on shutting off the  nebu-
lizer air supply the signal would decrease for a few seconds,
then suddenly rise to a larger value than was observed during
aerosolization and afterwards decrease to the background
level.  One possible explanation for this phenomenon is that
after aerosol generation was stopped, there remained sus-
pended in the air near the burner inlet a relatively large
proportion of the smaller droplets which were more easily
vaporized in the burner and more sensitively detected  than
the larger ones.  Alternatively, perhaps the signal increased
as the cooling of the flame by water lessened.

A tee was inserted between the atomization flask and the
burner inlet through which larger water droplets not elimin-
ated in the atomization flask might be drained and prevented
from entering the burner.  This modification improved  per-
                             13.

-------
formance and resulted in a detectability of 10 ppm triethyl-
phosphate.  It was observed that the burner's sample inlet
remained much drier after insertion of the tee.  Samples
containing 1 ppm TEP could be detected only occasionally.

The burner used was originally designed to measure phosphorus
or sulfur in air.  To achieve a rapid response and to keep
hydrogen consumption low the combustion chamber volume, sam-
ple inlet and hydrogen jet, which affects flame size, were
kept quite small.  It was thought that an increase in the
sample inlet crossectional area would allow more liquid
aerosol to reach the flame and that a larger hydrogen jet
would increase the size of the flame and also its tempera-
ture and would permit more droplets to vaporize and be mea-
sured.  These modifications were made and resulted in improved
sensitivity and samples containing 1 ppm TEP could be detected
consistently-  In some experiments, the sample inlet tube was
heated electrically.  At times it seemed that this pre-vol-
atilization step improved the phosphorus response.  Since
the benefit was not always obtainable this procedure was
abandoned.  If additional sensitivity is desired pre-heating
of the sample should be further investigated.

On substituting the ultrasonic nebulizer for the pneumatic
one, performance was significantly improved.  It became pos-
sible to detect 0.01 ppm TEP.  With the pneumatic nebulizer
1 ppm TEP was the lowest concentration that could be measured.
In addition, flame outs were practically eliminated.  With
the ultrasonic nebulizer, the aerosol particles are produced
at a lower pressure, essentially ambient, compared to at
least 6 psi when using the pneumatic device.  Further, fewer
large droplets and less liquid condensation were observed.
It is believed that operation at ambient pressure and/or
elimination of many of the large droplets greatly reduced
the number of flame-outs encountered.  Improved sensitivity
is also attributed to the smaller aerosol particle size.

The instrument response to 0.1 ppm TEP in water aerosolized
ultrasonically is shown in Fig. 3.  The response to water is
shown also for comparison.  The two spikes observable in the
TEP response tracing are believed to be caused by water drop-
lets in the instrument disturbing the air flow pattern.  Such
spikes are more frequently encountered when the aspirator is
not heated.  It is believed that raising the pneumatic system
to a higher temperature by heating the aspirator or not cool-
ing the burner to avoid water condensation will eliminate the
spikes.  The initial response could be observed in about 3
seconds and the response was 90 percent complete in about 15
seconds.
                             14.

-------
llJ
o
Q_
CO
UJ
o:
           O.I PPM  TER
             IN  H20
H20
                   TIME ( I"/MINUTE)



                       FIGURE  3


            RESPONSE  TO O.IPPM TER IN H20
                             15.

-------
In Figure 4 is shown the effect of concentration on response.
When plotted on graph paper a linear relationship was ob-
served between the net signal (instrument response to sample-
instrument response to water blank) and the phosphorus con-
centration in the sample.

As part of a study to determine the feasibility of disting-
uishing between organic and inorganic phosphorus compounds,
some measurements were made with ammonium phosphate.  A
smaller response was obtained with dibasic ammonium phosphate
than with triethylphosphate.  Using solutions containing 0.18
parts per million of phosphorus each, the average ratio of
the responses obtained with triethylphosphate and water was
1.7 while the average ratio of responses with ammonium phos-
phate and water was 1.25.  No attempts were made to ascertain
the reason for this difference in response between the two
phosphorus compounds, but it may be that the inorganic phos-
phorus compound is not as readily broken down as the organic
one at the relatively low flame temperature used.

Distinction Between Inorganic and Organic Phosphorus Compounds,

One of the purposes of the research was to determine the
feasibility of obtaining measurements of both the organic
and the inorganic phosphorus content of the water.  Since
the flame emission determination of phosphorus at the ana-
lytical wavelength of 525 nm depends on the excitation of
the POH radicals, the instrument should respond to the total
phosphorus content of the sample, both inorganic and organ-
ically bound.  A separation of the two types of phosphorus
compounds was, therefore, required in order to obtain mea-
surements of inorganic and organic phosphorus in the water.
The use of ion exchange resins for this purpose was studied.

Satisfactory results were achieved with a mixed ion exchange
resin, Rohm and Haas Amberlite Monobed MB-1 resin.  A one
foot long bed of the resin was used and samples were passed
only one time through the bed.  For these experiments diam-
monium hydrogen phosphate was used as the inorganic and
triethylphosphate as the organic phosphorus compound.  In
order to avoid interaction of the ion exchange resin with
elements other than phosphorus,  solutions were prepared with
water that was previously passed through the ion exchange
resin bed.  Typical results from flame emission measurements
are shown in the table below:

                 Sample                        Current (amps)
                                                       — 8
1 ppm TEP in deionized HO                       4.1x10
                                                       — 8
1 ppm TEP in deionized H20 after passage through 3.9x10
      resin bed.
                            16.

-------
40
0
        RESPONSE  ( AMPS X 10'*)
                  FIGURE 4
   RESPONSE - CONCENTRATION  RELATIONSHIP
            FOR TEP IN WATER
                       17.

-------
                 Sample

3 ppm  (NH )   HPO  in deionized H 0
         424               2
3 ppm  (NH )   HPO  in deionized H 0 after pas-   2.9x10
         424               2
Current (amps)
        -8
  3.7x10
    sage through resin bed
deionized H 0
  2.8x10
The data showed that removal of inorganic phosphorus by the
ion exchange resin was practically complete after a single
passage through the ion exchange resin bed.  The organic
phosphorus compound, on the other hand, was not removed by
the ion exchange resin.

To distinguish between inorganic and organic phosphorus one
would perform two measurements.  One sample would be mea-
sured directly for total phosphorus content.  A second
sample would be passed through a mixed ion exchange resin
bed for removal of inorganic phosphorus and a flame emis-
sion measurement of the effluent from the ion exchange
resin bed gives a measure of the organic phosphorus content.
The difference between the two measurements represents the
inorganic phosphorus.   This scheme requires that only phos-
phorus gives a measureable response or alternatively com-
pensation for, or prior removal of interfering substances.

It should be possible to get a measurement of inorganic and
organic phosphorus in the presence of interfering cation
concentrations by the use of two ion exchange columns.  The
sample would be passed through a cation exchange column to
replace cations in the sample by hydrogen ions.  A determi-
nation on part of the resultant solution would give a measure
of total phosphorus.  Another portion of the solution would
be passed through a second column containing an anion ex-
change resin to remove inorganic phosphorus.  A measurement
on the solution obtained from this second exchange procedure
would provide the organic phosphorus content.  As before,
the difference between the two measurements would give the
inorganic phosphorus content.

This hypothesis was tested with two different cation exchange
resins, J.  T. Baker No. 4622 and Rohm and Haas IR-120 used in
conjunction with J. T. Baker anion resin No. 4606.  It was
observed that a solution of dibasic ammonium phosphate con-
taining 0.5 ppm phosphorus gave the same response whether
passed through the cation exchange columns or not.  (The so-
lution was prepared with water that had previously been
passed through a cation exchange column.)  When subsequently
                              18.

-------
passed through an anion exchange  resin bed, the resultant
solution gave the same emission response as water passed
through an anion exchange  resin bed.  These data show that
the cation exchange  resin  does not affect inorganic phos-
phorus, and that the anion exchange bed removes the inor-
ganic phosphorus.

One unexpected result was  obtained.  Tap water gave an
emission response of lower intensity than water that was
passed through either of the two  cation exchange columns.
 (Identical signals were obtained  from tap water and water
passed through the anion exchange resin bed.)  It seems
that on passage through the cation exchange columns, some-
thing was introduced into  the sample that gives a response.
It is believed that  the response  is not due to hydrogen ions
obtained in the exchange reaction, but due to some other
material introduced  from the cation exchange resin bed.
Contact with Rohm and Haas indicates that this is possible
and that conditioning of the column prior to use is recom-
mended .

Potential Interferences.

Work done prior to this study with a number of inorganic
compounds showed that under the experimental conditions
used at that time, of those cations and anions tested,
only sodium and magnesium  gave a  measureable response.  The
concentrations used  corresponded  to the mean concentration
found in potable water supplies of the 50 largest U.S. cities,
as published by the  U.S. Geological Survey.

During this study, only the responses from calcium and sod-
ium compounds were measured.  Solutions of calcium chloride
and sodium nitrate were prepared  with water that was pre-
viously passed through the mixed  ion exchange resin bed.
Chloride and nitrate ions  are known not to respond at the
concentrations used.  Measurements were made with solutions
containing up to 170 ppm calcium.  The lowest calcium con-
centration that was  reliably detected was 32 ppm which gave
a signal corresponding to  about 3 ppb of phosphorus.  In the
concentration range  of 50  to 170  ppm calcium, the method was
found to be about 5000 times more sensitive towards phos-
phorus than towards  calcium.

Sensitivity of the instrument towards sodium was considerably
greater than towards calcium.  Measurements performed with
sodium concentrations between 7 and 112 ppm, indicated a
minimum detectability of 3 ppm of sodium.  The signal from
this concentration was equivalent to that from about 4 ppb
of phosphorus.  Over the concentration range measured instru-
ment sensitivity was 750 to 1000  times greater toward phos-
phorus than towards  sodium.
                             19.

-------
The possibility of removing interfering concentrations  of
sodium and calcium by ion exchange was also studied.  Solu-
tions containing 169 ppm calcium or 112 ppm sodium were
passed through a one foot bed of Rohm and Haas Amberlite
Monobed ion exchange resin MB-1, a mixed resin.  Flame
emission measurements were then made.  In both cases, the
emission intensities were identical to those obtained with
deionized water, indicating complete removal of detectable
sodium or calcium in one pass through the ion exchange
column.  For analysis of samples containing inorganic phos-
phorus, a cation exchanger should be used.

Design of Prototype Instrument.

On the basis of the experimental work described above,  the
prototype instrument to be delivered to EPA was designed.
The performance achieved in the experimental investigations
was judged to be satisfactory and the optical filter, photo-
multiplier tube, ignitor and with minor modification the
burner and pneumatic system were considered suitable for
use in the prototype.  Since a laboratory variable power
supply for the photomultiplier tube and Keithley electro-
meter for measuring the response were used in the investi-
gative phase appropriate substitutes were sought for incor-
poration into the prototype.  A power supply and an amplifier
were selected that NUCOR has used previously in radiation
monitors.

A portable cabinet, 12" x 14-1/8" x 18" with removable  front
panel was selected.  It was decided to mount all pneumatic
components needed to operate the instrument on the front
panel except for the compressor used to supply the air  re-
quired by the instrument.  Power supply and electrometer
were mounted on a base inside the cabinet.  Any sample  pre-
paration, e.g., aerosolization, ion exchange separation, etc.
would also be performed externally.

The burner, filter, and photomultiplier tube are mounted on
the bottom plate of a light tight housing.  A Baird-Atomic
Inc. one-inch diameter filter with peak wavelength at 525
nm (catalog number 11-97-2)  and an RCA 1P21 photomultiplier
tube are used.  A 1-1/2" long, 1/4" i.d. stainless steel
tube, threaded into the burner, extends through the burner
housing and front panel and serves as the inlet for the com-
bustion air and the sample.   The ignitor is mounted on  the
outside of the burner housing.

Air from a compressor, external to the instrument is supplied
through a fitting in the front panel.   A Cast Manufacturing
Co., Model 1531 oilless compressor is furnished.  Part  of the
air from'the compressor is vented before entering the instru-
                             20.

-------
ment.  In the instrument the air stream is split.  A portion
regulated with a Hoke "Milli-Mite" metering valve with micro-
meter provides the driving air for the aspirator which is
mounted on the bottom of the burner housing.  The rest of
the air is passed through a Dwyer Instruments, Inc., Model
VFB 4 inch flowmeter to the outside of the burner near the
flame and serves to cool the burner.  The aspirator, heated
with a Chromalox cartridge heater to prevent water condensa-
tion, attached to the exit of the burner creates a slight
negative pressure at the sample inlet.  This allows combus-
tion air and sample to be sucked into the burner.  The
combustion air flow rate is adjustable by means of the Hoke
valve.  Air flow control valve and flowmeter are mounted on
the front panel.  Metered hydrogen fuel from a cylinder is
supplied by means of a fitting on the instrument panel.

Electronic Design.

The electrical design of the equipment consists primarily of
a photomultiplier detector operated in the d-c mode operating
into a d-c electrometer amplifier with ameter readout.  The
circuitry in block diagram form is shown below, Figure 5.
The instrument wiring diagram is Figure 6.
                 ruse
                                        DC.
                         FIGURE 5.
                              21.

-------
KJ
                                                                                                        IGNITOR BOARD
                                                                                                                                                     ncM»i    1   HIGH VOLTAGE _
                                                                                                                                                                  POWER SUPPLY
                                                                                                                                                                   UOOCL HO.FIJ
                                                                                                                                                                 VEHUS  SCIENTIFIC INC
                                                                                    >iu««. MSTWJMCNT WIRING DIAGRAM

-------
The photomultiplier tube is operated at approximately 800
volts d-c as supplied by the H.V- power supply-  The output
of the photo tube is inserted directly to a high impedance
electrometer employing negative feedback for stability.  The
photomultiplier tube produces an electrical signal commen-
surate with the quantity of light which strikes the photo ,Q
cathode.  The electrical signal, approximately 10~^ to 10-
amperes depending on the light level is measured and indicated
by means of a hybrid-tube-transistor electrometer amplifier
employing negative feedback.  The electrometer operates from
a self-contained battery powered power converter.  The out-
put of the electrometer is displayed on a 4-1/2" panel meter.

The photomultiplier tube operates on 800 volts derived from
a sealed H.V. power supply module.  The high voltage power
supply is operated via a low voltage power supply operating
from normal line power.
                             23.

-------
                         SECTION V.

                        DISCUSSION
Comparison of the data obtained with phosphorus in air and
phosphorus in water samples  (Figures 2 and 4) indicates that
different phenomena control the emission response.  With
phosphorus in air samples, the response varies linearly with
concentration when plotted on log-log paper.  With phosphorus
in water samples, the response-concentration relationship is
linear.  Further, the response to phosphorus in air is much
larger than that to phosphorus in water.

For example, a net response of 1x10   amps was obtained with
0.03 ug/1 phosphorus in air and with 2.5 ppm TEP  (or 0.4 ppm
phosphorus) in water.  In the case of air samples, about
0.02 ug/min of phosphorus was introduced into the instrument
and except for small amounts  (losses on sample inlet wall,
etc.) was probably almost entirely detected.  Water was
aerosolized at a rate of about 3 cc/min. and phosphorus at
a rate of 1.2 ug/min.  Phosphorus in air was supplied
through a hypodermic needle inserted into the instrument's
sample inlet tube.  Consequently, phosphorus losses prior
to reaching the instrument were minimal.  The nebulizer hose,
which supplied the aerosol sample, was mounted about 3 inches
from the instrument sample inlet in order to minimize flame
outs through introduction of excessive amounts of water.
Significant amounts of aerosol emitted by the hose never
entered the instrument.  In addition, considerable quantities
of liquid that were aerosolized in the nebulizer chamber con-
densed in the 2 feet long sample hose.  Assuming that 25 per-
cent of the aerosolized sample reached the burner, phosphorus
in water passing through the burner at a rate of 0.3 ug/min.
gave a response equal to that from phosphorus in air passing
through the burner at a rate of only 0.021 ug/min.  Phos-
phorus in water is therefore detected much less sensitively
than in air.  It is believed that the poorer sensitivity
with water samples may be attributed to (a)   the need to
volatilize the phosphorus prior to excitation during the
short residence time in the flame, and  (b)   cooling of the
flame by water.  In separate experiments,  it was observed
that on cooling the burner, phosphorus in air is measured
at reduced sensitivity.  Preheating of the aerosol or re-
ducing air flow rate to increase residence time in the flame
may improve further the sensitivity to phosphorus in water.
                              24.

-------
                       SECTION VI

                    ACKNOWLEDGEMENTS
The NUCOR Project Manager was Mr. H. J. Cooley.  Dr. M. J.
Prager was Principal Investigator.  Mr. Cooley and Mr. F.
Riggin developed the electronic circuitry-  Mr. E. Weirich
assisted in the mechanical design of the instrument.

EPA Project Officer was Dr. T. B. Hoover, Southeast Environ-
mental Research Laboratory-
                              25.

-------
                       SECTION VII

                       REFERENCES
1.  Prager, M. J.  "Pollution Monitoring  by Flame Emission
    Spectroscopy", Optical Spectra,  5,  No.  8,  p.  28 (1971).

2.  Prager, M. J., Deblinger, B.,  and Kalinsky,  J. L.  "Water
    Monitoring for Trace Quantities  of  Organophosphorus
    Compounds with a Hydrogen Flame  Emission Detector" pre-
    sented at the 159th Meeting, American Chemical Society,
    Houston, Texas, February, 1970.
                                   4 U. S. GOVERNMENT PRINTING OFFICE : 1973—514-156/359
                            26.

-------
 SELECTED WATER
 RESOURCES ABSTRACTS
 INPUT TRANSACTION FORM
                  1. Report So.
                                                         3. Ai-c "i.5ion A'o
                                   w
 4. Title    Automated Water Monitoring  Instrument
          for Phosphorus  Contents
 7. Author(s)
             Prager, Manfred J.
 9. organization   NUCOR corporation, Nuclear Division
              2 Richwood Place, Denville,  N.J. 07834
                                   S.  Report Date

                                   6.

                                   8.  J rformi $ Organization
                                      Report No.
                                   10  Projt-ct No.

                                      16020-GSB
                                      Contract/Grant No.

                                      68-01-0111
                                      Typi  / Repa and
                                      Period Covered
 12. Sf -nsorin Organ vtion

 15. Supplementary Notes

    Environmental Protection Agency Report No. EPA*-R4-73~026,  June 1973
 16. Abstract  A prototype  instrument was developed by NUCOR Corporation,
Denville,  N.J., for automatically monitoring total phosphorus in water.
The analytical principle employed was  flame emission photometry.  Phos-
phorus  compounds burned in a hydrogen  flame emit at about 525 millimicrons
Conditions were established for the sensitive measurement of phosphorus  in
water.   Operating parameters investigated included fuel and air flow rates
burner  configuration,  operating temperature, method of  sample aerosoliza-
tion, etc.  Using an ultrasonic nebulizer to aerosolize samples of tri-
ethylphosphate in water,  it was possible to detect phosphorus at a concen-
tration of less than 2 parts per billion.   A procedure  was worked out  for
distinguishing between organic and inorganic phosphorus with ion exchange
resins.   In measurements designed to determine interference by sodium  and
calcium,  it was observed that the method is about 1000  times more sensi-
tive towards phosphorus than towards sodium and 5000 times more sensitive
towards phosphorus than towards calcium.   A prototype instrument was
designed,  fabricated,  tested, and delivered to EPA, Southeast Environmenta
Research Laboratory.
 17a. Descriptors
 17b. Identifiers
 17 c. COWRR Field & Group
 18. Availability
19.  Security Class.
   (Repoi )

V).  Se  rityC> .s.
21. ffo. of
   Pages

 3. Pr a
                                               Send To :
                                               WATER RESOURCES SCIENTIFIC INFORMATION CENTER
                                               U.S. DEPARTMENTOF THE INTERIOR
                                               WASHINGTON. D. C. 2O24O
 Abstractor
  'C. 102 :REV IUNF !P

-------