4500-SiO2 SILICA*

4500-SiO2 A. Introduction

1. Occurrence and Significance because one reagent fewer is used; one timing step is eliminated;
and many natural waters can be analyzed without dilution, which is
Silicon does not occur free in nature, but rather as free silica not often the case with Method D. Method D is recommended for
(SiO2) in coarsely crystalline (quartz, rock crystal, amethyst, the low range, from 0.04 to 2 mg SiO2/L. This range also can be
etc.) and microcrystalline (flint, chert, jasper, etc.) varieties of extended if necessary. Such extension may be desirable if interfer-
quartz, the major component of sand and sandstone. Silicon is ence is expected from tannin, color, or turbidity. A combination of
found in combination with other elements in silicates, repre- factors renders Methods D, E, and F less susceptible than Method C
sented by feldspar, hornblende, mica, asbestos, and other clay to those interferences; also, the blue color in Methods D, E, and F
minerals. Silicates also occur in rocks, such as granite, basalt, is more stable than the yellow color in Method C. However, many
and shale. Silicon therefore is usually reported as silica (SiO2) samples will require dilution because of the high sensitivity of the
when rocks, sediments, soils, and water are analyzed. The aver- method. Permanent artificial color standards are not available for the
age abundance of silica in different rock types is 7 to 80%, in blue color developed in Method D.
typical soils 50 to 80%, and in surface and groundwater 14 mg/L. The yellow color produced by Method C and the blue color
The common aqueous forms of silica are H4SiO4 and produced by Methods D, E, and F are affected by high concentra-
H3SiO4⫺. In the presence of magnesium, it can form scale tions of salts. With seawater the yellow color intensity is decreased
deposits in boilers and in steam turbines. It is considered a by 20 to 35% and the blue color intensity is increased by 10 to 15%.
nonessential trace element for most plants, but essential for most When waters of high ionic strength are analyzed by these methods,
animals. Chronic exposure to silica dust can be toxic. There is no use silica standards of approximately the same ionic strengths.2
U.S. EPA drinking water standard MCL for silica. Method E or F may be used where large numbers of samples
A more complete discussion of the occurrence and chemistry are analyzed regularly. Method 3111D is recommended for
of silica in natural waters is available.1 broad-range use. Although Method 3111D is usable from 1 to
300 mg SiO2/L, optimal results are obtained from about 20 to
2. Selection of Method 300 mg/L. The range can be extended upward by dilution if
necessary. This method is rapid and does not require any timing step.
Perform analyses by the electrothermal atomic absorption The inductively coupled plasma method (3120) also may be
method (3113B) or one of the colorimetric methods (C, D, E, or F), used in analyses for silica.
depending on the fraction to be measured. The inductively coupled
plasma mass spectrometric method (3125) or the inductively cou-
3. Sampling and Storage
pled plasma method (3120) also may be applied successfully in
most cases (with lower detection limits), even though silica is not Collect samples in bottles of polyethylene, other plastic, or
specifically listed as an analyte in the method. hard rubber, especially if there will be a delay between collection
Methods 3120 and 3125 determine total silica. Methods C, D, and analysis. Borosilicate glass is less desirable, particularly
E, and F determine molybdate-reactive silica. As noted in 4500- with waters of pH above 8 or with seawater, in which cases a
SiO2.C.4, it is possible to convert other forms of silica to the significant amount of silica in the glass can dissolve. Freezing to
molybdate-reactive form for determination by these methods. preserve samples for analysis of other constituents can lower
Method 3111D determines more than one form of silica. It will soluble silica values by as much as 20 to 40% in waters that have
determine all dissolved silica and some colloidally dispersed a pH below 6. Do not acidify samples for preservation because
silica. The determination of silica present in micrometer and silica precipitates in acidic solutions.
submicrometer particles will depend on the size distribution,
composition, and structure of the particles; thus Method 3111D 4. References
cannot be said to determine total silica.
Method C is recommended for relatively pure waters containing 1. HEM, J.D. 1985. Study and interpretation of the Chemical Character-
from 0.4 to 25 mg SiO2/L. As with most colorimetric methods, the istics of Natural Water, 3rd ed. U.S. Geol. Surv. Water Supply Pap.
range can be extended, if necessary, by diluting, by concentrating, No. 2254.
or by varying the light path. Interferences due to tannin, color, and 2. FANNING, K.A. & M.E.Q. PILSON. 1973. On the spectrophotometric
turbidity are more severe with this method than with Method D. determination of dissolved silica in natural waters. Anal. Chem.
Moreover, the yellow color produced by Method C has a limited 45:136.
stability and attention to timing is necessary. When applicable,
however, it offers greater speed and simplicity than Method D 5. Bibliography

ROY, C.J. 1945. Silica in natural waters. Amer. J. Sci. 243:393.

* Approved by Standard Methods Committee, 1997. Editorial revisions, 2011.
VAIL, J.G. 1952. The Soluble Silicates, Their Properties and Uses.
Joint Task Group: 20th Edition—(4500-SiO2.F)—Scott Stieg (chair), Bradford R. Reinhold Publishing Corp., New York, N.Y. Vol. 1, pp. 95–97,
Fisher, Owen B. Mathre, Theresa M. Wright. 100 –161.

1
 SILICA (4500-SiO2)/Molybdosilicate Method

4500-SiO2 B. (Reserved)*

4500-SiO2 C. Molybdosilicate Method

1. General Discussion TABLE 4500-SiO2.I. SELECTION OF LIGHT PATH LENGTH FOR VARIOUS
SILICA CONCENTRATIONS
a. Principle: Ammonium molybdate at pH approximately Method D
1.2 reacts with silica and any phosphate present to produce
heteropoly acids. Oxalic acid is added to destroy the molyb- Silica in 55 mL Final Volume
dophosphoric acid but not the molybdosilicic acid. Even if Method C ␮g
phosphate is known to be absent, the addition of oxalic acid Light Silica in 55 mL
is highly desirable and is a mandatory step in both this method Path Final Volume 650 nm 815 nm
and Method D. The intensity of the yellow color is propor- cm ␮g Wavelength Wavelength
tional to the concentration of “molybdate-reactive” silica. In 1 200–1300 40–300 20–100
at least one of its forms, silica does not react with molybdate 2 100–700 20–150 10–50
even though it is capable of passing through filter paper and 5 40–250 7–50 4–20
is not noticeably turbid. It is not known to what extent such 10 20–130 4–30 2–10
“unreactive” silica occurs in waters. Terms such as “colloi-
dal,” “crystalloidal,” and “ionic” have been used to distin-
guish among various forms of silica but such terminology
cannot be substantiated. “Molybdate-unreactive” silica can be a. Sodium bicarbonate, NaHCO3, powder.
converted to the “molybdate-reactive” form by heating or b. Sulfuric acid, H2SO4, 1N.
fusing with alkali. Molybdate-reactive or unreactive does not c. Hydrochloric acid, HCl, 1 ⫹ 1.
imply reactivity, or lack of it, toward other reagents or pro- d. Ammonium molybdate reagent: Dissolve 10 g
cesses. (NH4)6Mo7O24 䡠 4H2O in water, with stirring and gentle warm-
b. Interference: Because both apparatus and reagents may ing, and dilute to 100 mL. Filter if necessary. Adjust to pH 7 to
contribute silica, avoid using glassware as much as possible 8 with silica-free NH4OH or NaOH and store in a polyethylene
and use reagents low in silica. Also, make a blank determi- bottle to stabilize. (If the pH is not adjusted, a precipitate
nation to correct for silica so introduced. In both this method gradually forms. If the solution is stored in glass, silica may
and Method D, tannin, large amounts of iron, color, turbidity, leach out and cause high blanks.) If necessary, prepare silica-free
sulfide, and phosphate interfere. Treatment with oxalic acid NH4OH by passing gaseous NH3 into distilled water contained in
eliminates interference from phosphate and decreases inter- a plastic bottle.
ference from tannin. If necessary, use photometric compen- e. Oxalic acid solution: Dissolve 7.5 g H2C2O4 䡠 H2O in
sation to cancel interference from color or turbidity. water and dilute to 100 mL.
c. Minimum detectable concentration: Approximately 1 mg f. Stock silica solution: Dissolve 4.73 g sodium metasilicate
SiO2/L can be detected in 50-mL nessler tubes. nonahydrate, Na2SiO3 䡠 9H2O, in water and dilute to 1000 mL.
d. Quality control (QC): The QC practices considered to be an For work of highest accuracy, analyze 100.0-mL portions by the
integral part of each method are summarized in Table 4020:I. gravimetric method.1 Store in a tightly stoppered plastic bottle.
g. Standard silica solution: Dilute 10.00 mL stock solution to
2. Apparatus 1000 mL with water; 1.00 mL ⫽ 10.0 ␮g SiO2. Calculate exact
concentration from concentration of stock silica solution. Store
a. Platinum dishes, 100-mL. in a tightly stoppered plastic bottle.
b. Colorimetric equipment: One of the following is required: h. Permanent color solutions:
1) Spectrophotometer, for use at 410 nm, providing a light 1) Potassium chromate solution: Dissolve 630 mg K2CrO4 in
path of 1 cm or longer. See Table 4500-SiO2:I for light path water and dilute to 1 L.
selection. 2) Borax solution: Dissolve 10 g sodium borate decahydrate,
2) Filter photometer, providing a light path of 1 cm or longer Na2B4O7 䡠 10H2O, in water and dilute to 1 L.
and equipped with a violet filter having maximum transmittance
near 410 nm. 4. Procedure
3) Nessler tubes, matched, 50-mL, tall form.
a. Color development: To 50.0 mL sample add in rapid suc-
3. Reagents cession 1.0 mL 1 ⫹ 1 HCl and 2.0 mL ammonium molybdate
reagent. Mix by inverting at least six times and let stand for 5 to
For best results, set aside and use batches of chemicals low in 10 min. Add 2.0 mL oxalic acid solution and mix thoroughly.
silica. Use distilled reagent water in making reagents and dilu- Read color after 2 min but before 15 min, measuring time from
tions. Store all reagents in plastic containers to guard against addition of oxalic acid. Because the yellow color obeys Beer’s
high blanks. law, measure photometrically or visually.

2
 SILICA (4500-SiO2)/Molybdosilicate Method

TABLE 4500-SIO2:II. PREPARATION OF PERMANENT COLOR STANDARDS correctness of these permanent artificial standards by comparing
FOR VISUAL DETERMINATION OF SILICA them visually against standards prepared by analyzing portions
Potassium of the standard silica solution. Use permanent artificial color
Values Chromate Borax standards only for visual comparison.
in Silica Solution Solution Water
␮g mL mL mL 5. Calculation
0 0.0 25 30
100 1.0 25 29 ␮g SiO2 (in 55 mL final volume)
200 2.0 25 28 mg SiO2/L ⫽
400 4.0 25 26 mL sample
500 5.0 25 25
750 7.5 25 22 Report whether NaHCO3 digestion was used.
1000 10.0 25 20
6. Precision and Bias

A synthetic sample containing 5.0 mg SiO2/L, 10 mg Cl⫺/L,

b. To detect the presence of molybdate-unreactive silica, di- 0.20 mg NH3-N/L, 1.0 mg NO3⫺-N/L, 1.5 mg organic N/L, and
gest sample with NaHCO3 before color development. This di- 10.0 mg PO43⫺/L in distilled water was analyzed in 19 labora-
gestion is not necessarily sufficient to convert all molybdate- tories by the molybdosilicate method with a relative standard
unreactive silica to the molybdate-reactive form. Complex sili- deviation of 14.3% and a relative error of 7.8%.
cates and higher silica polymers may require extended fusion Another synthetic sample containing 15.0 mg SiO2/L, 200 mg
with alkali at high temperatures or digestion under pressure for Cl⫺/L, 0.800 mg NH3-N/L, 1.0 mg NO3⫺-N/L, 0.800 mg organic
complete conversion. Omit digestion if all the silica is known to N/L, and 5.0 mg PO43⫺/L in distilled water was analyzed in 19
react with molybdate. laboratories by the molybdosilicate method, with a relative stan-
Prepare a clear sample by filtration if necessary. Place 50.0 mL, dard deviation of 8.4% and a relative error of 4.2%.
or a smaller portion diluted to 50 mL, in a 100-mL platinum dish. A third synthetic sample containing 30.0 mg SiO2/L, 400 mg
Add 200 mg silica-free NaHCO3 and digest on a steam bath for Cl⫺/L, 1.50 mg NH3-N/L, 1.0 mg NO3⫺-N/L, 0.200 mg organic
1 h. Cool and add slowly, with stirring, 2.4 mL 1N H2SO4. Do N/L, and 0.500 mg PO43⫺/L, in distilled water was analyzed in
not interrupt analysis but proceed at once with remaining steps. 20 laboratories by the molybdosilicate method, with a relative
Transfer quantitatively to a 50-mL nessler tube and make up to standard deviation of 7.7% and a relative error of 9.8%.
mark with water. (Tall-form 50-mL nessler tubes are convenient All results were obtained after sample digestion with
for mixing even if the solution subsequently is transferred to an NaHCO3.
absorption cell for photometric measurement.)
c. Preparation of standards: If NaHCO3 pretreatment is used,
add to the standards (approximately 45 mL total volume) 200 mg 7. Reference
NaHCO3 and 2.4 mL 1N H2SO4, to compensate both for the
slight amount of silica introduced by the reagents and for the 1. EATON, A.D., L.S. CLESCERI & A.E. GREENBERG, eds. 1995. Standard
Methods for the Examination of Water and Wastewater., 19th ed.
effect of the salt on color intensity. Dilute to 50.0 mL.
American Public Health Assoc., American Water Works Assoc., &
d. Correction for color or turbidity: Prepare a special blank Water Environment Fed., Washington, D.C.
for every sample that needs such correction. Carry two identical
portions of each such sample through the procedure, including
NaHCO3 treatment if this is used. To one portion, add all 8. Bibliography
reagents as directed in ¶ a above. To the other portion, add HCl
and oxalic acid but no molybdate. Adjust photometer to zero DIENERT, F. & F. WANDENBULCKE. 1923. On the determination of silica in
waters. Bull. Soc. Chim. France 33:1131, Compt. Rend. 176:1478.
absorbance with the blank containing no molybdate before read-
DIENERT, F. & F. WANDENBULCKE. 1924. A study of colloidal silica.
ing absorbance of molybdate-treated sample. Compt. Rend. 178:564.
e. Photometric measurement: Prepare a calibration curve SWANK, H.W. & M.G. MELLON. 1934. Colorimetric standards for silica.
from a series of approximately six standards to cover the opti- Ind. Eng. Chem., Anal. Ed. 6:348.
mum ranges cited in Table 4500-SiO2:I. Follow direction of ¶ a TOURKY, A.R. & D.H. BANGHAM. 1936. Colloidal silica in natural waters
above on suitable portions of standard silica solution diluted to and the “silicomolybdate” colour test. Nature 138:587.
50.0 mL in nessler tubes. Set photometer at zero absorbance with BIRNBAUM, N. & G.H. WALDEN. 1938. Co-precipitation of ammonium
water and read all standards, including a reagent blank, against silicomolybdate and ammonium phosphomolybdate. J. Amer.
water. Plot micrograms silica in the final (55 mL) developed Chem. Soc. 60:66.
solution against photometer readings. Run a reagent blank and at KAHLER, H.L. 1941. Determination of soluble silica in water: A photo-
metric method. Ind. Eng. Chem., Anal. Ed. 13:536.
least one standard with each group of samples to confirm that the
NOLL, C.A. & J.J. MAGUIRE. 1942. Effect of container on soluble silica
calibration curve previously established has not shifted. content of water samples. Ind. Eng. Chem., Anal. Ed. 14:569.
f. Visual comparison: Make a set of permanent artificial color SCHWARTZ, M.C. 1942. Photometric determination of silica in the pres-
standards, using K2CrO4 and borax solutions. Mix liquid vol- ence of phosphates. Ind. Eng. Chem., Anal. Ed. 14:893.
umes specified in Table 4500-SiO2:II and place them in well- GUTTER, H. 1945. Influence of pH on the composition and physical
stoppered, appropriately labeled 50-mL nessler tubes. Verify aspects of the ammonium molybdates. Compt. Rend. 220:146.

3
 SILICA (4500-SiO2)/Heteropoly Blue Method

MILTON, R.F. 1951. Formation of silicomolybdate. Analyst 76:431. STRICKLAND, J.D.H. 1952. The preparation and properties of silico mo-
KILLEFFER, D.H. & A. LINZ. 1952. Molybdenum Compounds, Their lybdic acid. J. Amer. Chem. Soc. 74:862, 868, 872.
Chemistry and Technology. Interscience Publishers, New York, CHOW, D.T.W. & R.J. ROBINSON. 1953. The forms of silicate available
N.Y. pp. 1–2, 42– 45, 67– 82, 87–92. for colorimetric determination. Anal. Chem. 25:646.

4500-SiO2 D. Heteropoly Blue Method

1. General Discussion that produces reagents that are dark even when freshly pre-
pared.*
a. Principle: The principles outlined under 4500-SiO2.C.1a,
also apply to this method. The yellow molybdosilicic acid is 4. Procedure
reduced by means of aminonaphtholsulfonic acid to heteropoly a. Color development: Proceed as in 4500-SiO2.C.4a up to
blue. The blue color is more intense than the yellow color of and including the words, “Add 2.0 mL oxalic acid solution and
4500-SiO2.C provides increased sensitivity. mix thoroughly.” Measuring time from the moment of adding
b. Interference: See 4500-SiO2.C.1b. oxalic acid, wait at least 2 min but not more than 15 min, add
c. Minimum detectable concentration: Approximately 2.0 mL reducing agent, and mix thoroughly. After 5 min, mea-
20 ␮g SiO2/L can be detected in 50-mL nessler tubes and sure blue color photometrically or visually. If NaHCO3 pretreat-
50 ␮g SiO2/L spectrophotometrically with a 1-cm light path at ment is used, follow 4500-SiO2.C.4b.
815 nm. b. Photometric measurement: Prepare a calibration curve from a
d. Quality control (QC): The QC practices considered to be series of approximately six standards to cover the optimum range
an integral part of each method are summarized in Table indicated in Table 4500-SiO2:I. Carry out the steps described above
4020:I. on suitable portions of standard silica solution diluted to 50.0 mL in
nessler tubes; pretreat standards if NaHCO3 digestion is used (see
2. Apparatus 4500-SiO2.C.4b). Adjust photometer to zero absorbance with dis-
tilled water and read all standards, including a reagent blank, against
a. Platinum dishes, 100-mL. distilled water. If necessary to correct for color or turbidity in a
b. Colorimetric equipment: One of the following is required: sample, see 4500-SiO2.C.4d. To the special blank add HCl and
1) Spectrophotometer, for use at approximately 815 nm. The oxalic acid, but no molybdate or reducing agent. Plot micrograms
color system also obeys Beer’s law at 650 nm, with appreciably silica in the final 55 mL developed solution against absorbance. Run
reduced sensitivity. Use light path of 1 cm or longer. See Table a reagent blank and at least one standard with each group of samples
4500-SiO2:I for light path selection. to check the calibration curve.
2) Filter photometer, provided with a red filter exhibiting c. Visual comparison: Prepare a series of not less than 12
maximum transmittance in the wavelength range of 600 to standards, covering the range 0 to 120 ␮g SiO2, by placing the
815 nm. Sensitivity improves with increasing wavelength. Use calculated volumes of standard silica solution in 50-mL nessler
light path of 1 cm or longer. tubes, diluting to mark with distilled water, and developing color
3) Nessler tubes, matched, 50-mL, tall form. as described in ¶ a above.

5. Calculation
3. Reagents
␮g SiO2 (in 55 mL final volume)
mg SiO2/L ⫽
For best results, set aside and use batches of chemicals low in mL sample
silica. Store all reagents in plastic containers to guard against
high blanks. Use distilled water that does not contain detectable Report whether NaHCO3 digestion was used.
silica after storage in glass.
All of the reagents listed in 4500-SiO2.C.3 are required, and in 6. Precision and Bias
addition: A synthetic sample containing 5.0 mg SiO2/L, 10 mg Cl⫺/L,
Reducing agent: Dissolve 500 mg 1-amino-2-naphthol-4- 0.200 mg NH3-N/L, 1.0 mg NO3⫺-N/L, 1.5 mg organic N/L, and
sulfonic acid and 1 g Na2SO3 in 50 mL distilled water, with 10.0 mg PO43⫺/L in distilled water was analyzed in 11 labora-
gentle warming if necessary; add this to a solution of 30 g tories by the heteropoly blue method, with a relative standard
NaHSO3 in 150 mL distilled water. Filter into a plastic bottle. deviation of 27.2% and a relative error of 3.0%.
Discard when solution becomes dark. Prolong reagent life by
storing in a refrigerator and away from light. Do not use
aminonaphtholsulfonic acid that is incompletely soluble or * Eastman No. 360 has been found satisfactory.

4
 SILICA (4500-SiO2)/Automated Method for Molybdate-Reactive Silica

A second synthetic sample containing 15 mg SiO2/L, 7. Bibliography

200 mg Cl⫺/L, 0.800 mg NH3-N/L, 1.0 mg NO3⫺-N/L, 0.800
mg organic N/L, and 5.0 mg PO43⫺/L in distilled water was BUNTING, W.E. 1944. Determination of soluble silica in very low con-
analyzed in 11 laboratories by the heteropoly blue method, centrations. Ind. Eng. Chem., Anal. Ed. 16:612.
with a relative standard deviation of 18.0% and a relative STRAUB, F.G. & H. GRABOWSKI. 1944. Photometric determination of
error of 2.9%. silica in condensed steam in the presence of phosphates. Ind. Eng.
Chem., Anal. Ed. 16:574.
A third synthetic sample containing 30.0 mg SiO2/L, 400 mg
BOLTZ, D.F. & M.G. MELLON. 1947. Determination of phosphorus,
Cl⫺/L, 1.50 mg NH3-N/L, 1.0 mg NO3⫺-N/L, 0.200 mg organic germanium, silicon, and arsenic by the heteropoly blue method. Ind.
N/L, and 0.500 mg PO43⫺/L in distilled water was analyzed in 10 Eng. Chem., Anal. Ed. 19:873.
laboratories by the heteropoly blue method with a relative stan- MILTON, R.F. 1951. Estimation of silica in water. J. Appl. Chem. (Lon-
dard deviation of 4.9% and a relative error of 5.1%. don) 1:(Supplement No. 2) 126.
All results were obtained after sample digestion with CARLSON, A.B. & C.V. BANKS. 1952. Spectrophotometric determination
NaHCO3. of silicon. Anal. Chem. 24:472.

4500-SiO2 E. Automated Method for Molybdate-Reactive Silica

1. General Discussion

a. Principle: This method is an adaptation of the heteropoly

blue method (Method D) using a continuous-flow analytical
instrument.
b. Interferences: See 4500-SiO2.C.1b. If particulate matter is
present, filter sample or use a continuous filter as an integral part
of the system.
c. Application: This method is applicable to potable, surface,
domestic, and other waters containing 0 to 20 mg SiO2/L. The
range of concentration can be broadened to 0 to 80 mg/L by
substituting a 15-mm flow cell for the 50-mm flow cell shown in
Figure 4500-SiO2:1.
d. Quality control (QC): The QC practices considered to be an
integral part of each method are summarized in Table 4020:I.

2. Apparatus

Automated analytical equipment: An example of the continu-

ous-flow analytical instrument consists of the interchangeable Figure 4500-SiO2:1. Silica manifold.
components shown in Figure 4500-SiO2:1.

3. Reagents
4. Procedure
a. Sulfuric acid, H2SO4, 0.05M (0.1N).
b. Ammonium molybdate reagent: Dissolve 10 g Set up manifold as shown in Figure 4500-SiO2:1 and follow
(NH4)6Mo7O24 䡠 4H2O in 1 L 0.05M H2SO4. Filter and store in the general procedure described by the manufacturer. Determine
an amber plastic bottle. absorbance at 660 nm. Use quality control procedures given in
c. Oxalic acid solution: Dissolve 50 g oxalic acid in 900 mL Section 4020.
distilled water and dilute to 1 L.
d. Reducing agent: Dissolve 120 g NaHSO3 and 4 g Na2SO3 5. Calculation
in 800 mL warm distilled water. Add 2 g 1-amino-2-naphthol-
4-sulfonic acid, mix well, and dilute to 1 L. Filter into amber
plastic bottle for storage. Prepare standard curves by plotting response of standards
To prepare working reagent, dilute 100 mL to 1 L with processed through the manifold against SiO2 concentration in
distilled water. Make working reagent daily. standards. Compute sample SiO2 concentration by comparing
e. Standard silica solution: See 4500-SiO2.C.3g. sample response with standard curve.

5
 SILICA (4500-SiO2)/Flow Injection Analysis for Molybdate-Reactive Silicate

6. Precision and Bias the coefficient of variation (95% confidence level at 7.1 mg/L)
was 1.6%. For 0 to 80 mg SiO2/L, when a 15-mm flow cell
For 0 to 20 mg SiO2/L, when a 50-mm flow cell was used was used at 50 samples/h, detection limit was 0.5 mg/L,
at 40 samples/h, the detection limit was 0.1 mg/L, sensitivity sensitivity was 31 mg/ L, and coefficient of variation at
(concentration giving 0.398 absorbance) was 7.1 mg/L, and 31 mg/L was 1.5%.

4500-SiO2 F. Flow Injection Analysis for Molybdate-Reactive Silicate

1. General Discussion

Silicate reacts with molybdate under acidic conditions to form

yellow beta-molybdosilicic acid. This acid is subsequently re-
duced with stannous chloride to form a heteropoly blue complex
that has an absorbance maximum at 820 nm. Oxalic acid is added
to reduce the interference from phosphate.
Collect samples in polyethylene or other plastic bottles and
refrigerate at 4°C. Chemical preservation for silica is not rec-
ommended. Adding acid may cause polymerization of reactive
silicate species. Freezing decreases silicate concentrations, espe-
cially at concentrations greater than 100 ␮g SiO2/L. If filtration Figure 4500-SiO2:2. FIA manifold.
is required, preferably use a 0.45-␮m TFE filter. Samples may be
held for 28 d.
Also see 4500-SiO2.A, D, and E, and Section 4130, Flow
d. Absorbance detector, 820 nm, 10-nm bandpass.
Injection Analysis (FIA).
e. Injection valve control and data acquisition system.
Interferences: Remove large or fibrous particulates by filtering
sample through inert filter. 3. Reagents
The interference due to phosphates is reduced by the addition
of oxalic acid as a reagent on the flow injection manifold. By the
Use reagent water (⬎10 megohm) to prepare carrier and all
following method, a solution of 1000 ␮g P/L was determined as
solutions. To prevent bubble formation, degas carrier and buffer
20 ␮g SiO2/L. Verify extent of phosphate interference by deter-
with helium. Pass He at 140 kPa (20 psi) through a helium
mining a solution of phosphate at the highest concentration that
degassing tube. Bubble He through 1 L solution for 1 min. All
is expected to occur.
reagents can also be prepared on a weight/volume basis if
Tannin and large amounts of iron or sulfides are interferences.
desired.
Remove sulfides by boiling an acidified sample. Add disodium
a. Molybdate: To a tared 500-mL container, add 20.0 g am-
EDTA to eliminate interference due to iron. Treat with oxalic
monium molybdate tetrahydrate [(NH4)6Mo7O24 䡠 4H2O]. Add
acid to decrease interference from tannin.
486 g warm water and 14.8 g conc sulfuric acid (H2SO4). Stir or
Sample color and turbidity can interfere. Determine presence
shake until dissolved. Store in plastic and refrigerate. Prepare
of these interferences by analyzing samples without the presence
fresh monthly and discard if precipitate or blue color is observed.
of the molybdate.
b. Oxalic acid: To a tared 500-mL container, add 50.0 g
Avoid silica contamination by storing samples, standards, and
oxalic acid (HO2CCO2H 䡠 2H2O) and 490 g water. Stir or shake
reagents in plastic. Do not use glass-distilled water for reagents
until dissolved. Store in plastic.
or standards.
c. Stannous chloride: To a tared 1-L container, add 978 g
water. Add 40.0 g conc H2SO4. Dissolve 2.0 g hydroxylamine
2. Apparatus hydrochloride in this solution. Then dissolve 0.30 g stannous
chloride. Prepare fresh weekly.
Flow injection analysis equipment consisting of: d. Stock silicate standard, 100 mg SiO2/L: In a 1-L volumet-
a. FIA injection valve with sample loop or equivalent. ric flask dissolve 0.473 g sodium metasilicate nonahydrate
b. Multichannel proportioning pump. (Na2SiO3 䡠 9H2O) in approximately 800 mL water. Dilute to
c. FIA manifold (Figure 4500-SiO2:2) with tubing heater and mark and invert three times. Alternatively, use a commercially
flow cell. Relative flow rates only are shown in Figure 4500- prepared standard solution, especially if nonstoichiometry of the
SiO2:2. Tubing volumes are given as an example only; they may solid metasilicate is suspected; the original degree of polymer-
be scaled down proportionally. Use manifold tubing of an inert ization of the sodium metasilicate, which depends on storage
material, such as TFE.* time, can affect free silica concentration of the resulting solution.
e. Standard silicate solutions: Prepare silicate standards in the
desired concentration range, using the stock standard (¶ d
* Teflon, or equivalent. above), and diluting with water.

6
 SILICA (4500-SiO2)/Flow Injection Analysis for Molybdate-Reactive Silicate

4. Procedure method.1 These gave a mean of 4.86 ␮g SiO2/L, a standard

deviation of 0.31 ␮g SiO2/L, and an MDL of 0.78 ␮g SiO2/L.
Set up a manifold equivalent to that in Figure 4500-SiO2:2 and
follow method supplied by manufacturer or laboratory standard
operating procedure. 7. Quality Control

5. Calculations The QC practices considered to be an integral part of each

Prepare standard curves by plotting absorbance of standards method are summarized in Table 4020:I.
processed through the manifold versus silicate concentration.
The calibration curve is linear.
8. Reference
6. Precision and Bias
1. U.S. ENVIRONMENTAL PROTECTION AGENCY. 1989. Definition and Pro-
Twenty-one replicates of a 5.0-␮g SiO2/L standard were an- cedure for the Determination of Method Detection Limits. Appendix
alyzed with a 780-␮L sample loop by a published MDL B to 40 CFR 136 rev. 1.11 amended June 30, 1986. 49 CFR 43430.