4500-F⫺ FLUORIDE*

4500-F⫺ A. Introduction

Fluoride may occur naturally in water or it may be added in TABLE 4500F⫺:I. CONCENTRATION OF SOME SUBSTANCES CAUSING
controlled amounts. Some fluorosis may occur when the flu- 0.1-MG/L ERROR AT 1.0 MG F⫺/L IN FLUORIDE METHODS
oride level exceeds the recommended limits. In rare instances, Method C Method D
the naturally occurring fluoride concentration may approach (Electrode) (SPADNS)
10 mg/L.
Accurate determination of fluoride has increased in impor- Conc Type of Conc Type of
Substance mg/L Error* mg/L Error*
tance with the growth of the practice of fluoridation of water
supplies as a public health measure. Maintenance of an optimal Alkalinity 7 000 ⫹ 5 000 ⫺
fluoride concentration is essential in maintaining effectiveness (CaCO3)
and safety of the fluoridation procedure. Aluminum (Al3⫹) 3.0 ⫺ 0.1† ⫺
Chloride (Cl⫺) 20 000 7 000 ⫹
1. Preliminary Treatment Chlorine 5 000 Remove
completely
with
Among the methods suggested for determining fluoride ion
arsenite
(F⫺) in water, the electrode and colorimetric methods are the Color & turbidity Remove or
most satisfactory. Because both methods are subject to errors due compensate
to interfering ions (Table 4500:F⫺:I), it may be necessary to for
distill the sample as directed in Section 4500-F⫺.B before mak- Iron 200 ⫺ 10 ⫺
ing the determination. When interfering ions are not present in Hexametaphosphate 50 000 1.0 ⫹
excess of the tolerance of the method, the fluoride determination ([NaPO3]6)
may be made directly without distillation. Phosphate 50 000 16 ⫹
(PO43⫺)
2. Selection of Method Sulfate 50 000 ⫺ 200 ⫺
(SO42⫺)
The electrode methods (C and G) are suitable for fluoride * ⫹ denotes positive error
concentrations from 0.1 to more than 10 mg/L. Adding the ⫺ denotes negative error
Blank denotes no measurable error.
prescribed buffer frees the electrode method from most interfer- † On immediate reading. Tolerance increases with time: after 2 h, 3.0; after 4 h,
ences that adversely affect the SPADNS colorimetric method 30.
and necessitate preliminary distillation. Some substances in in-
dustrial wastes, such as fluoborates, may be sufficiently concen-
trated to present problems in electrode measurements and will
not be measured without a preliminary distillation. Fluoride
measurements can be made with an ion-selective electrode and Ion chromatography (Section 4110) is an acceptable method if
either an expanded-scale pH meter or a specific ion meter, weaker eluents are used to separate fluoride from interfering
usually without distillation, in the time necessary for electrode peaks or fluoride can be determined by capillary ion electropho-
equilibration. resis (Section 4140).
The SPADNs method (D) has a linear analytical range of 0 to The flow injection method (G) is a convenient automated
1.40 mg F⫺/L. Use of a nonlinear calibration can extend the technique for analyzing large numbers of samples.
range to 3.5 mg F⫺/L. Color development is virtually instanta-
neous. Color determinations are made photometrically, using 3. Sampling and Storage
either a filter photometer or a spectrophotometer. A curve de-
veloped from standards is used for determining the fluoride Preferably use polyethylene bottles for collecting and storing
concentration of a sample. samples for fluoride analysis. Glass bottles are satisfactory if
Fluoride also may be determined by the automated complex- previously they have not contained high-fluoride solutions. Al-
one method, Method E. ways rinse bottle with a portion of sample.
For the SPADNs method, never use an excess of dechlorinat-
ing agent. Dechlorinate with sodium arsenite rather than sodium
* Approved by Standard Methods Committee, 1997. Editorial revisions, 2011.
Joint Task Group: 20th Edition (4500-F⫺.G)—Scott Stieg (chair), Bradford R. thiosulfate when using the SPADNS method because the latter
Fisher, Owen B. Mathre, Theresa M. Wright. may produce turbidity that causes erroneous readings.

1
 FLUORIDE (4500-F⫺)/Preliminary Distillation Step

4500-F⫺ B. Preliminary Distillation Step

1. Discussion

Fluoride can be separated from other nonvolatile constituents

in water by conversion to hydrofluoric or fluosilicic acid and
subsequent distillation. The conversion is accomplished by using
a strong, high-boiling acid. To protect against glassware etching,
hydrofluoric acid is converted to fluosilicic acid by using soft
glass beads. Quantitative fluoride recovery is approached by
using a relatively large sample. Acid and sulfate carryover are
minimized by distilling over a controlled temperature range.
Distillation will separate fluoride from most water samples.
Some tightly bound fluoride, such as that in biological materials,
may require digestion before distillation, but water samples
seldom require such drastic treatment. Distillation produces a
distillate volume equal to that of the original water sample, so
usually it is not necessary to incorporate a dilution factor when
expressing analytical results. The distillate will be essentially
free of substances that might interfere with the fluoride determi-
nation if the apparatus used is adequate and distillation has been
carried out properly. The only common volatile constituent
likely to cause interference with colorimetric analysis of the
distillate is chloride. When the concentration of chloride is high
enough to interfere, add silver sulfate to the sulfuric acid distill-
ing mixture to minimize the volatilization of hydrogen chloride.
CAUTION: Heating an acid–water mixture can be hazard-
ous if precautions are not taken: Mix acid and water thor-
oughly before heating. Use of a quartz heating mantle and a
magnetic stirrer in the distillation apparatus simplifies the mixing
step. Figure 4500-Fⴚ:1. Direct distillation apparatus for fluoride.

2. Apparatus The preferred apparatus is illustrated in Figure 4500-F⫺:1.

b. Quartz hemispherical heating mantle, for full-voltage op-
a. Distillation apparatus consisting of a 1-L round-bottom eration.
long-neck borosilicate glass boiling flask, a connecting tube, an c. Magnetic stirrer, with TFE-coated stirring bar.
efficient condenser, a thermometer adapter, and a thermometer d. Soft glass beads.
that can be read to 200°C. Use standard taper joints for all
connections in the direct vapor path. Position the thermometer so 3. Reagents
the bulb always is immersed in boiling mixture. The apparatus
should be disassembled easily to permit adding sample. Substi- a. Sulfuric acid, H2SO4, conc, reagent grade.
tuting a thermoregulator and necessary circuitry for the ther- b. Silver sulfate, Ag2SO4, crystals, reagent grade.
mometer is acceptable and provides some automation.
Alternative types of distillation apparatus may be used. Carefully 4. Procedure
evaluate any apparatus for fluoride recovery and sulfate carryover.
The critical points are obstructions in the vapor path and trapping of a. Place 400 mL distilled water in the distilling flask and, with
liquid in the adapter and condenser. (The condenser should have a the magnetic stirrer operating, carefully add 200 mL conc
vapor path with minimum obstruction. A double-jacketed con- H2SO4. Keep stirrer in operation throughout distillation. Add a
denser, with cooling water in the outer jacket and the inner spiral few glass beads and connect the apparatus as shown in Figure
tube, is ideal, but other condensers are acceptable if they have 4500-F⫺:1, making sure all joints are tight. Begin heating and
minimum obstructions. Avoid using Graham-type condensers.) continue until flask contents reach 180°C (because of heat re-
Avoid using an open flame as a heat source if possible, because heat tention by the mantle, it is necessary to discontinue heating when
applied to the boiling flask above the liquid level causes superheat- the temperature reaches 178°C to prevent overheating ). Discard
ing of vapor and subsequent sulfate carryover. distillate. This process removes fluoride contamination and ad-
CAUTION: Regardless of apparatus used, provide for thorough justs the acid–water ratio for subsequent distillations.
mixing of sample and acid; heating a non-homogenous acid– b. After the acid mixture remaining in the steps outlined in ¶ a
water mixture will result in bumping or possibly a violent above, or previous distillations, has cooled to 80°C or below, add
explosion. 300 mL sample, with stirrer operating, and distill until the

2
 FLUORIDE (4500-F⫺)/Ion-Selective Electrode Method

temperature reaches 180°C. To prevent sulfate carryover, turn distillation. After periods of inactivity, similarly flush still and
off heat before 178°C. Retain the distillate for analysis. discard distillate.
c. Add Ag2SO4 to the distilling flask at the rate of 5 mg/mg
Cl⫺ when the chloride concentration is high enough to interfere 5. Interpretation of Results
(see Table 4500-F⫺:I). The recovery of fluoride is quantitative within the accuracy of
d. Use H2SO4 solution in the flask repeatedly until contaminants the methods used for its measurement.
from samples accumulate to such an extent that recovery is affected
or interferences appear in the distillate. Check acid suitability peri- 6. Bibliography
odically by distilling standard fluoride samples and analyzing for
both fluoride and sulfate. After distilling samples containing more BELLACK, E. 1958. Simplified fluoride distillation method. J. Amer.
Water Works Assoc. 50:530.
than 3 mg F⫺/L, flush still by adding 300 mL distilled water, BELLACK, E. 1961. Automatic fluoride distillation. J. Amer. Water Works
redistill, and combine the two fluoride distillates. If necessary, Assoc. 53:98.
repeat flushing until the fluoride content of the last distillate is at a ZEHNPFENNIG, R.G. 1976. Letter to the editor. Environ. Sci. Technol.
minimum. Include additional fluoride recovered with that of the first 10:1049.

4500-F⫺ C. Ion-Selective Electrode Method

1. General Discussion ion concentration is greater than one-tenth the concentration of

fluoride ion. At the pH maintained by the buffer, no hydroxide
a. Principle: The fluoride electrode is an ion-selective sensor. interference occurs.
The key element in the fluoride electrode is the laser-type doped Fluoborates are widely used in industrial processes. Dilute
lanthanum fluoride crystal across which a potential is established solutions of fluoborate or fluoboric acid hydrolyze to liberate
by fluoride solutions of different concentrations. The crystal fluoride ion but in concentrated solutions, as in electroplating
contacts the sample solution at one face and an internal reference wastes, hydrolysis does not occur completely. Distill such sam-
solution at the other. The cell may be represented by: ples or measure fluoborate with a fluoborate-selective electrode.
Also distill the sample if the dissolved solids concentration
AgAgCl, Cl⫺(0.3M), F⫺(0.001M) LaF3 test exceeds 10 000 mg/L.
c. Quality control (QC): The QC practices considered to be an
solutionreference electrode integral part of each method are summarized in Table 4020:I.

The fluoride electrode can be used with a standard calomel 2. Apparatus

reference electrode and almost any modern pH meter having an
expanded millivolt scale. Calomel electrodes contain both me- a. Expanded-scale or digital pH meter or ion-selective meter.
tallic and dissolved mercury; therefore, dispose of them only in b. Sleeve-type reference electrode: Do not use fiber-tip refer-
approved sites or recycle. For this reason, the Ag/AgCl reference ence electrodes because they exhibit erratic behavior in very
electrode is preferred. dilute solutions.
The fluoride electrode measures the ion activity of fluoride in c. Fluoride electrode.
solution rather than concentration. Fluoride ion activity depends d. Magnetic stirrer, with TFE-coated stirring bar.
on the solution total ionic strength and pH, and on fluoride e. Timer.
complexing species. Adding an appropriate buffer provides a
nearly uniform ionic strength background, adjusts pH, and 3. Reagents
breaks up complexes so, in effect, the electrode measures con-
centration. a. Stock fluoride solution: Dissolve 221.0 mg anhydrous so-
b. Interference: Table 4500-F⫺:I lists common interferences. dium fluoride, NaF, in distilled water and dilute to 1000 mL;
Fluoride forms complexes with several polyvalent cations, no- 1.00 mL ⫽ 100 ␮g F⫺.
tably aluminum and iron. The extent to which complexation b. Standard fluoride solution: Dilute 100 mL stock fluoride
takes place depends on solution pH, relative levels of fluoride, solution to 1000 mL with distilled water; 1.00 mL ⫽ 10.0 ␮g F⫺.
and complexing species. However, CDTA (cyclohexylenedi- c. Fluoride buffer: Place approximately 500 mL distilled water
aminetetraacetic acid), a component of the buffer, preferentially in a 1-L beaker and add 57 mL glacial acetic acid, 58 g NaCl,
will complex interfering cations and release free fluoride ions. and 4.0 g 1,2 cyclohexylenediaminetetraacetic acid (CDTA).*
Concentrations of aluminum, the most common interference, up Stir to dissolve. Place beaker in a cool water bath and add slowly
to 3.0 mg/L can be complexed preferentially. In acid solution, F⫺ 6N NaOH (about 125 mL) with stirring, until pH is between 5.0
forms a poorly ionized HF 䡠 HF complex but the buffer main- and 5.5. Transfer to a 1-L volumetric flask and add distilled
tains a pH above 5 to minimize hydrogen fluoride complex
formation. In alkaline solution, hydroxide ion also can interfere
with electrode response to fluoride ion whenever the hydroxide * Also known as 1,2 cyclohexylenedinitrilotetraacetic acid.

3
 FLUORIDE (4500-F⫺)/SPADNS Method

water to the mark. This buffer, as well as a more concentrated cycle semilogarithmic graph paper. Plot milligrams F⫺ per liter
version, is available commercially. In using the concentrated on the logarithmic axis (ordinate), with the lowest concentration
buffer, follow the manufacturer’s directions. at the bottom of the graph. Plot millivolts on the abscissa. From
the potential measurement for each sample, read the correspond-
4. Procedure ing fluoride concentration from the standard curve.
The known-additions method may be substituted for the cal-
a. Instrument calibration: No major adjustment of any instru- ibration method described. Follow the directions of the instru-
ment normally is required to use electrodes in the range of 0.2 to ment manufacturer.
2.0 mg F⫺/L. For those instruments with zero at center scale, Selective-ion meters may necessitate using a slightly altered
adjust calibration control so the 1.0 mg F⫺/L standard reads at procedure, such as preparing 1.00 and 10.0 mg F⫺/L standards or
the center zero (100 mV) when the meter is in the expanded- some other concentration. Follow the manufacturer’s directions.
scale position. This cannot be done on some meters that do not Commercial standards, often already diluted with buffer, fre-
have a millivolt calibration control. To use a selective-ion meter, quently are supplied with the meter. Verify the stated fluoride
follow the manufacturer’s instructions. concentration of these standards by comparing them with stan-
b. Preparation of fluoride standards: Prepare a series of stan- dards prepared by the analyst.
dards by diluting with distilled water 5.0, 10.0, and 20.0 mL of
standard fluoride solution to 100 mL with distilled water. These 5. Calculation
standards are equivalent to 0.5, 1.0, and 2.0 mg F⫺/L. ␮g F⫺
c. Treatment of standards and sample: In 100-mL beakers or mg F⫺/L ⫽
mL sample
other convenient containers, add by volumetric pipet from 10 to
25 mL standard or sample. Bring standards and sample to same 6. Precision and Bias
temperature, preferably room temperature. Add an equal volume
of buffer. The total volume should be sufficient to immerse the A synthetic sample containing 0.850 mg F⫺/L in distilled
electrodes and permit operation of the stirring bar. water was analyzed in 111 laboratories by the electrode method,
d. Measurement with electrode: Immerse electrodes in each of with a relative standard deviation of 3.6% and a relative error of
the fluoride standard solutions and measure developed potential 0.7%.
while stirring on a magnetic stirrer. Avoid stirring before im- A second synthetic sample containing 0.750 mg F⫺/L, 2.5 mg
mersing electrodes because entrapped air around the crystal can (NaPO3)6/L, and 300 mg alkalinity/L added as NaHCO3, was
produce erroneous readings or needle fluctuations. Let electrodes analyzed in 111 laboratories by the electrode method, with a
remain in the solution 3 min (or until reading is constant) before relative standard deviation of 4.8% and a relative error of 0.2%.
taking a final millivolt reading. A layer of insulating material A third synthetic sample containing 0.900 mg F⫺/L, 0.500 mg
between stirrer and beaker minimizes solution heating. With- Al/L, and 200 mg SO42⫺/L was analyzed in 13 laboratories by
draw electrodes, rinse with distilled water, and blot dry between the electrode method, with a relative standard deviation of 2.9%
readings. (CAUTION: Blotting may poison electrode if not done and a relative error of 4.9%.
gently.) Repeat measurements with samples.
When using an expanded-scale pH meter or selective-ion 7. Bibliography
meter, frequently recalibrate the electrode by checking potential FRANT, M.S. & J.W. ROSS, JR. 1968. Use of total ionic strength adjust-
reading of the 1.00-mg F⫺/L standard and adjusting the calibra- ment buffer for electrode determination of fluoride in water sup-
tion control, if necessary, until meter reads as before. plies. Anal. Chem. 40:1169.
If a direct-reading instrument is not used, plot potential mea- HARWOOD, J.E. 1969. The use of an ion-selective electrode for routine
surement of fluoride standards against concentration on two- analysis of water samples. Water Res. 3:273.

4500-F⫺ D. SPADNS Method

1. General Discussion made almost instantaneous. Under such conditions, however, the
effect of various ions differs from that in the conventional
a. Principle: The SPADNS colorimetric method is based on the alizarin methods. The selection of dye for this rapid fluoride
reaction between fluoride and a zirconium-dye lake. Fluoride reacts method is governed largely by the resulting tolerance to these
with the dye lake, dissociating a portion of it into a colorless ions.
complex anion (ZrF62⫺); and the dye. As the amount of fluoride b. Interference: Table 4500-F⫺:I lists common interferences.
increases, the color produced becomes progressively lighter. Because these are neither linear in effect nor algebraically addi-
The reaction rate between fluoride and zirconium ions is tive, mathematical compensation is impossible. Whenever any
influenced greatly by the acidity of the reaction mixture. If the one substance is present in sufficient quantity to produce an error
proportion of acid in the reagent is increased, the reaction can be of 0.1 mg/L or whenever the total interfering effect is in doubt,

4
 FLUORIDE (4500-F⫺)/SPADNS Method

distill the sample. Also distill colored or turbid samples. In some reference, set photometer at some convenient point (0.300 or
instances, sample dilution or adding appropriate amounts of 0.500 absorbance) with the prepared 0 mg F⫺/L standard.
interfering substances to the standards may be used to compen- b. Sample pretreatment: If the sample contains residual chlo-
sate for the interference effect. If alkalinity is the only significant rine, remove it by adding 1 drop (0.05 mL) NaAsO2 solution/
interference, neutralize it with either hydrochloric or nitric acid. 0.1 mg residual chlorine and mix. (Sodium arsenite concentrations
Chlorine interferes and provision for its removal is made. of 1300 mg/L produce an error of 0.1 mg/L at 1.0 mg F⫺/L.)
Volumetric measurement of sample and reagent is extremely c. Color development: Use a 50.0-mL sample or a portion
important to analytical accuracy. Use samples and standards at diluted to 50 mL with distilled water. Adjust sample temperature
the same temperature or at least within 2°C. Maintain constant to that used for the standard curve. Add 5.00 mL each of
temperature throughout the color development period. Prepare SPADNS solution and zirconyl-acid reagent, or 10.00 mL acid-
different calibration curves for different temperature ranges. zirconyl-SPADNS reagent; mix well and read absorbance, first
c. Quality control (QC): The QC practices considered to be an setting the reference point of the photometer as above. If the
integral part of each method are summarized in Table 4020:I. absorbance falls beyond the range of the standard curve, repeat
using a diluted sample.
2. Apparatus
5. Calculation
Colorimetric equipment: One of the following is required:
a. Spectrophotometer, for use at 570 nm, providing a light path A B
of at least 1 cm. mg F⫺/L ⫽ ⫻
mL sample C
b. Filter photometer, providing a light path of at least 1 cm and
equipped with a greenish yellow filter having maximum trans- where:
mittance at 550 to 580 nm. A ⫽ ␮g F⫺ determined from plotted curve,
B ⫽ final volume of diluted sample, mL, and
3. Reagents C ⫽ volume of diluted sample used for color development, mL.

a. Standard fluoride solution: Prepare as directed in the elec- When the prepared 0 mg F⫺/L standard is used to set the
trode method, 4500-F⫺.C.3b. photometer, alternatively calculate fluoride concentration as fol-
b. SPADNS solution: Dissolve 958 mg SPADNS, sodium lows:
2-(parasulfophenylazo)-1,8-dihydroxy-3,6-naphthalene disul-
fonate, also called 4,5-dihydroxy-3-(parasulfophenylazo)-2,7- A0 ⫺ Ax
naphthalenedisulfonic acid trisodium salt, in distilled water and mg F⫺/L ⫽
A0 ⫺ A1
dilute to 500 mL. This solution is stable for at least 1 year if
protected from direct sunlight. where:
c. Zirconyl-acid reagent: Dissolve 133 mg zirconyl chloride A0 ⫽ absorbance of the prepared 0 mg F⫺/L standard,
octahydrate, ZrOCl2 䡠 8H2O, in about 25 mL distilled water. A1 ⫽ absorbance of a prepared 1.0 mg F⫺/L standard, and
Add 350 mL conc HCl and dilute to 500 mL with distilled water. Ax ⫽ absorbance of the prepared sample.
d. Acid zirconyl-SPADNS reagent: Mix equal volumes of
SPADNS solution and zirconyl-acid reagent. The combined re-
6. Precision and Bias
agent is stable for at least 2 years.
e. Reference solution: Add 10 mL SPADNS solution to 100 mL A synthetic sample containing 0.830 mg F⫺/L and no interfer-
distilled water. Dilute 7 mL conc HCl to 10 mL and add to the ence in distilled water was analyzed in 53 laboratories by the
diluted SPADNS solution. The resulting solution, used for setting SPADNS method, with a relative standard deviation of 8.0% and a
the instrument reference point (zero), is stable for at least 1 year. relative error of 1.2%. After direct distillation of the sample, the
Alternatively, use a prepared standard of 0 mg F⫺/L as a reference. relative standard deviation was 11.0% and the relative error 2.4%.
f. Sodium arsenite solution: Dissolve 5.0 g NaAsO2 and dilute A synthetic sample containing 0.570 mg F⫺/L, 10 mg Al/L,
to 1 L with distilled water. (CAUTION: Toxic—avoid ingestion.) 200 mg SO42⫺/L, and 300 mg total alkalinity/L was analyzed in
53 laboratories by the SPADNS method without distillation, with
4. Procedure a relative standard deviation of 16.2% and a relative error of
7.0%. After direct distillation of the sample, the relative standard
a. Preparation of standard curve: Prepare fluoride standards deviation was 17.2% and the relative error 5.3%.
in the range of 0 to 1.40 mg F⫺/L by diluting appropriate A synthetic sample containing 0.680 mg F⫺/L, 2 mg Al/L,
quantities of standard fluoride solution to 50 mL with distilled 2.5 mg (NaPO3)6/L, 200 mg SO42⫺/L, and 300 mg total alka-
water. Pipet 5.00 mL each of SPADNS solution and zirconyl- linity/L was analyzed in 53 laboratories by direct distillation and
acid reagent, or 10.00 mL mixed acid-zirconyl-SPADNS re- SPADNS methods with a relative standard deviation of 2.8% and
agent, to each standard and mix well. Avoid contamination. Set a relative error of 5.9%.
photometer to zero absorbance with the reference solution and
obtain absorbance readings of standards. Plot a curve of the 7. Bibliography
milligrams fluoride-absorbance relationship. Prepare a new stan-
dard curve whenever a fresh reagent is made or a different BELLACK, E. & P.J. SCHOUBOE. 1968. Rapid photometric determination of
standard temperature is desired. As an alternative to using a fluoride with SPADNS-zirconium lake. Anal. Chem. 30:2032.

5
 FLUORIDE (4500-F⫺)/Complexone Method

4500-F⫺ E. Complexone Method

1. General Discussion

a. Principle: The sample is distilled in the automated system, and

the distillate is reacted with alizarin fluorine blue-lanthanum reagent
to form a blue complex that is measured colorimetrically at 620 nm.
b. Interferences: Interferences normally associated with the
determination of fluoride are removed by distillation.
c. Application: This method is applicable to potable, surface,
and saline waters as well as domestic and industrial wastewaters.
The range of the method, which can be modified by using the
adjustable colorimeter, is 0.1 to 2.0 mg F⫺/L.
d. Quality control (QC): The QC practices considered to be an
integral part of each method are summarized in Table 4020:I.

2. Apparatus

An example of the required continuous-flow analytical instru-

ment consists of the interchangeable components in the number
and manner indicated in Figure 4500-F⫺:2.

3. Reagents

a. Standard fluoride solution: Prepare in appropriate con-

centrations from 0.10 to 2.0 mg F⫺/L using the stock fluoride
solution (see 4500-F⫺.C.3a). Figure 4500-Fⴚ:2. Fluoride manifold.
b. Distillation reagent: Add 50 mL conc H2SO4 to about 600 mL
distilled water. Add 10.00 mL stock fluoride solution (see Section
4500-F⫺.C.3a; 1.00 mL ⫽ 100 ␮g F⫺) and dilute to 1000 mL.
c. Acetate buffer solution: Dissolve 60 g anhydrous sodium Set up manifold as shown in Figure 4500-F⫺:2 and follow the
acetate, NaC2H3O2, in about 600 mL distilled water. Add manufacturer’s instructions.
100 mL conc (glacial) acetic acid and dilute to 1 L.
d. Alizarin fluorine blue stock solution: Add 960 mg aliz-
5. Calculation
arin fluorine,* C14H7O4 䡠 CH2N(CH2 䡠 COOH)2, to 100 mL
distilled water. Add 2 mL conc NH4OH and mix until dye is
Prepare standard curves by plotting response of standards
dissolved. Add 2 mL conc (glacial) acetic acid, dilute to
processed through the manifold against constituent concentra-
250 mL and store in an amber bottle in the refrigerator.
tions in standards. Compute sample concentrations by comparing
e. Lanthanum nitrate stock solution: Dissolve 1.08 g La(NO3)3
sample response with standard curve.
in about 100 mL distilled water, dilute to 250 mL, and store in
refrigerator.
6. Precision and Bias
f. Working color reagent: Mix in the following order: 300 mL
acetate buffer solution, 150 mL acetone, 50 mL tertiary butanol,
In a single laboratory, four samples of natural water containing
36 mL alizarin fluorine blue stock solution, 40 mL lanthanum
from 0.40 to 0.82 mg F⫺/L were analyzed in septuplicate.
nitrate stock solution, and 2 mL polyoxyethylene 23 lauryl
Average precision was ⫾ 0.03 mg F⫺/L. To two of the samples,
ether.† Dilute to 1 L with distilled water. This reagent is stable
additions of 0.20 and 0.80 mg F⫺/L were made. Average recov-
for 2 to 4 d.
ery of the additions was 98%.
4. Procedure
7. Bibliography
No special handling or preparation of sample is required.
WEINSTEIN, L.H., R.H. MANDL, D.C. MCCUNE, J.S. JACOBSON & A.E.
HITCHCOCK. 1963. A semi-automated method for the determina-
* J.T. Baker Catalog No. J-112, or equivalent. tion of fluorine in air and plant tissues. Boyce Thompson Inst.
† Brij-35, available from ICI Americas, Wilmington, DE, or equivalent. 22:207.

6
 FLUORIDE (4500-F⫺)/Ion-Selective Electrode FIA

4500-F⫺ F. (Reserved)

4500-F⫺ G. Ion-Selective Electrode Flow Injection Analysis

1. General Discussion

a. Principle: Fluoride is determined potentiometrically by using

a combination fluoride-selective electrode in a flow cell. The fluo-
ride electrode consists of a lanthanum fluoride crystal across which
a potential is developed by fluoride ions. The reference cell is a
Ag/AgCl/Cl⫺ cell. The reference junction is of the annular liquid-
junction type and encloses the fluoride-sensitive crystal.
Also see 4500-F⫺.C and Section 4130, Flow Injection Anal-
ysis (FIA).
b. Interferences: Remove large or fibrous particulates by filtering
sample through glass wool. Guard against contamination from
reagents, water, glassware, and the sample preservation process.
The polyvalent cations Si4⫹, Al3⫹, and Fe3⫹ interfere by forming
complexes with fluoride. As part of the buffer reagent, 1,2-cyclo-
Figure 4500-Fⴚ:3. FIA fluoride manifold.
hexyldiaminetetraacetic acid (CDTA) is added to preferentially
complex these cations and eliminate this interference when these
concentrations do not exceed 3.0 mg Al3⫹/L and 20 mg Fe3⫹/L. or stir to mix thoroughly. Store fluoride electrode in this solution
Some interferents are removed by distillation; see 4500-F⫺.B. when it is not in use.
Drinking water samples generally do not require sample distil- d. Stock fluoride standard, 100.0 mg F⫺/L: In a 1-L volumet-
lation. ric flask, dissolve 0.2210 g sodium fluoride, NaF, in approxi-
c. Quality control (QC): The QC practices considered to be an mately 950 mL water. Dilute to mark with water and mix well.
integral part of each method are summarized in Table 4020:I. Store in a polyethylene bottle.
e. Standard fluoride solutions: Prepare fluoride standards in
2. Apparatus the desired concentration range, using the stock standard (¶ d
above), and diluting with water. A blank or zero concentration
Flow injection analysis equipment consisting of: standard cannot be prepared for this method because it will give
a. FIA injection valve with sample loop or equivalent. an undefined response from the fluoride electrode.
b. Multichannel proportioning pump.
c. FIA manifold (Figure 4500-F⫺:3) with tubing heater and 4. Procedure
ion-selective electrode flow cell. In Figure 4500-F⫺:3, relative
flow rates only are shown. Tubing volumes are given as an Set up a manifold equivalent to that in Figure 4500-F⫺:3 and
example only; they may be scaled down proportionally. Use follow method supplied by manufacturer or laboratory standard
manifold tubing of an inert material, such as TFE. operating procedure for this method.
d. Combination ion-selective electrode.
e. Injection valve control and data acquisition system. 5. Calculations

3. Reagents Prepare standard curves by plotting the electrode response to

standards processed through the manifold vs. fluoride concen-
Use reagent water (⬎10 megohm) for all solutions. To prevent tration. Standards greater than 1.0 mg F⫺/L will give positive
bubble formation, degas carrier and buffer with helium. Pass He peaks, standards less than 1.0 mg F⫺/L will give negative peaks,
at 140 kPa (20 psi) through a helium degassing tube. Bubble He and the 1.0 mg F⫺/L standard having the same concentration as
through 1 L solution for 1 min. the carrier will give no peak. The calibration curve gives a good
a. Carrier, 1.0 mg F⫺/L: Add 10 mL or 10 g stock fluoride fit to a second-order polynomial.
standard (¶ d below) to 990 mL water and mix well. It is not necessary to plot the response versus log[F⫺]; if this
b. Buffer: To a tared 1-L polyethylene container add 929.5 g is done, the calibration curve will still be a second-order poly-
water, 59.8 g glacial acetic acid, 30.0 g sodium hydroxide, nomial because there is a concentration-dependent kinetic effect
NaOH, 58.0 g sodium chloride, NaCl, 0.5 g stock fluoride in the flowing stream electrode system.
standard (¶ d below), and 4.0 g 1,2-cyclohexyldiaminetetraacetic
acid (CDTA) (also called trans-1,2-diaminocyclohexane). Stir 6. Precision and Bias
on a magnetic stir plate until all material has dissolved.
c. Electrode conditioning solution: To a tared 1-L container, The samples used in the studies described below were not
add 534 g buffer (¶ b above) and 500 g carrier (¶ a above). Shake distilled.

7
 FLUORIDE (4500-F⫺)/Ion-Selective Electrode FIA

TABLE 4500-F⫺:II. RESULTS OF SINGLE-LABORATORY STUDIES WITH a. Recovery and relative standard deviation: The results of
SELECTED MATRICES single-laboratory studies with various matrices are given in
Relative Table 4500-F⫺:II.
Known Standard b. MDL: A 390-␮L sample loop was used in the method
Sample/Blank Addition Recovery Deviation described above. Ten replicates of a 1.0-mg F⫺/L standard were
Matrix Designation mg F⫺/L % % run to obtain an MDL of 0.02 mg F⫺/L.
Wastewater Reference — 101 — c. Precision: Ten replicate standards of 2.0 mg F⫺/L gave a
treatment sample* % RSD of 0.5%.
plant Blank† 1.0 91 —
influent 2.0 97 —
Site A‡ 0 — 4.8
1.0 93 —
2.0 82 —
Site B‡ 0 — 6.4
1.0 96 —
2.0 86 —
Site C‡§ 0 — 15
1.0 99 —
2.0 86 —
Wastewater Reference — 103 —
treatment sample*
plant Blank† 1.0 97 —
effluent 2.0 97 —
Site A‡ 0 — ND
1.0 ND —
2.0 ND —
Site B‡ 0 — ⬍0.1
1.0 80 —
2.0 78 —
Site C‡ 0 — ⬍0.1
1.0 93 —
2.0 91 —
Landfill Reference — 99 —
leachate㛳 sample*
Blank† 1.0 87 —
2.0 88 —
Site A‡ 0 — 13
1.0 74 —
2.0 68 —
Site B‡ 0 — 10
1.0 68 —
2.0 73 —
Site C‡ 0 — 32
1.0 66 —
2.0 79 —
ND ⫽ not detectable.
* U.S. EPA QC sample, 1.81 mg F⫺/L.
† Determined in duplicate.
‡ Samples without known additions determined four times; samples with known
additions determined in duplicate. Typical difference between duplicates for
influent 5%, for effluent 6%.
§ Mean concentration 0.18 mg F⫺/L.
㛳 All sites had mean concentration of ⬍0.2 mg F⫺/L.