a.» P.
‘3“ n; m L .,..P..J ”P:

RILUW .w‘ . 211.. ..
a... .. av... : d $3...

«7" ”5:“
F
-_’\:
I V S
‘3?
J1
.t-
I;-
(qt
:3;
J

- Oupnx. . .I
. .\ an ARV : :- . a”.
x .3 r
n _ . t A. . J. u.
WV. . u .3. l QoJ\ _ JR”... A ”.3
cm-.. 9%...“ can . L.
' o. .wo..~

 

THESIS

JIIIT WWWHTIW I IIIH'fllel‘l‘lfiflfllH'IflT|1fiilffi| «

3 1293 10692 4180

‘ . v“ .‘M
. L:
. . .
..

  
   
 

 
 

BINDING BY

"BAG 8: SDNS‘

ABSTRACT

METHODS OF ANALYSIS FOR SULFUR IN SOILS

_ By

Ignacio Hernan Salcedo

A bibliographic review (including more than 70 references) of
available methods for soil and plant analysis for sulfur (S) was made.

Based on a criterion of sensitivity, accuracy, and simplicity,
three methods for total and extractable S, and extractable sulfate (son)
were selected .

The method used for total S consisted of an alkaline oxidation of
the soil samples with sodium hypobromite which transformed all 8 form to
son. The determination procedure was based on SO” reduction to H23
using a reducing mixture composed of HI, HOAc, and NaHzPOZ. The H25
generated was then measured colorimetrically by the methylene blue
method.

Based on results of preliminary tests, it was possible to modify
this procedure, thus reducing significantly the time of analysis and, in
one soil sample, increasing the amount of S measured.

Three Michigan soils -- Morley, Hoytville, and Miami series --
were used in this study. Total S was measured in natural soil samples
and in samples with 100 ppm of added S which were subsequently

incubated for two weeks at 20 C. Original S contents were 222, uzu, and

201 ppm of S, and the percent recoveries of added S were 100, 103.5, and

Ignacio Hernan Salcedo

and 106.2, respectively.

Extraction with water and with a 500 ppm P solution was used to
evaluate extractable S; this was measured with the methylene blue
method. These extractions were made on soil samples which had 0 and
6 weeks of incubation time and with 0 and 100 ppm of added S.
Significant increases in the amounts of extractable S due to incubation
were found with both extractants in the samples without added S. Only
one sample, with 100 ppm of added S, showed this same effect when
extracted with water.

These same extracts, after being clarified with FeCl3 or activated
charcoal depending on their degree of turbidity, were used for 804
analysis.

The method initially selected consisted of a precipitation of the
son with Ba(103)2. Colorimetry of the free IO3 after its reduction to

12' gave an indirect measurement of the 804 concentration.

As this method offered several problems, a new one was developed
using the same precipitation technique. Although not standardized in a
definite way, as little as 0.15 ppm of 804-8 was measured by the

colorimetric measurement of a iodine/acetonitrile system.

METHODS OF ANALYSIS FOR SULFUR IN SOILS

By

Ignacio Hernan Salcedo

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Crop and Soil Sciences

197a

This work is dedicated

to the memory of my father

and to my beloved families who,
by my side and from abroad,

helped me so much.

11

ACKNOWIEDGMENT

I wish to sincerely thank my major professor, Dr. Boyd G. Ellis,
for all of his thoughtful advise throughout the course of my studies
and research at this University, and for the correction of this
manuscript. I extend my acknowledgment to those members of the
Department of Crop and Soil Sciences who, in one way or another, helped
me in this achievment.

To Ing. Agr. Ichiro Mizuno, who suggested me to begin studies in
soil sulfur and supported them, goes my sincere appreciation.

I am also grateful to the University of Buenos Aires for the
financial support which has enabled me to complete this study.

And last, but not least, I wish to thank Liz Shields for her
professional help, and even more, for her friendship and support in

this undertaking.

iii

TABLE OF CONTENTS

INTRODUCTION 0 o o o o o n o o o c o o o o o o O o O O o o o o
erRATIJRE REVIEH o o o 0 o o o o o o o o o o o o o o o o o I

A. Total Soil S Extraction . . . . . . . . . . . . . . .

1. Reduction Methods . . . . . . . . . . . . . . . .
2 I oxfl-da'tion MethOd-S l O I I O O 0 O I O O O O O O O

B. Total Plant S Extraction . . . . . . . . . . . . . .
C. Extraction of Available S . . . . . . .

D. Quantitative Determination of

Turbidimetry .
Nephelometry .

{rknrula

Polarography .

Coulometry . .
Volumetry . . .

00 CD\} O\U’\

1...:

MATERIAIS AND METHODS. . . .

A. Materials . . . . .
B. Methods . . . . . .

. Total S in Soils

UIF'UNH

RESULTS AND DISCUSSION . . .

1. Total S Determination .
2. Extractable S Determination .

3. Extractable SO“ Determination

SUMMARY AND CONCLUSIONS . .

BIBLIOGRAPHY . . . . . . . .

iv

Absorption Spectrometry . . .
Atomic Absorption Spectrometry
Emission Spectrometry . .
Potentiometric Titrations . .

. Preparation of the Samples
. Extraction Procedures .

. Sulfur Determination by The
. Sulfate Determination by Iodine

Extracted

X-ray Emission SpectrOgraphy

m o

Methylene

Blue Method

Colorimetry . . .

LIST OF TABLES

Soil type, chemical and physical propoerties of the three

Michigan soils used . . . . . . . . . . . . . . . . . .

Influence of the time of reduction-distillation and of
residue dilution upon total soil-S determination . . .

Analysis of variance of the experiment of the influence
the reduction-distillation time and of residue dilution

Total soil-S and percent recovery of added S . . . . .
Hater-extractable S . . . . . . . . . . . . . . . . . .

Extractable S with a 500 ppm P solution . . . . . . . .

Page

18

27

28
31
35
36

LIST OF FIGURES

115332 35.139
1. Relationship between absorbance and S concentration . . . 43

2. Absorption spectra of the iodine/acetonitrile system . . 44

vi

INTRODUCTION

Sulfur, although an essential nutrient for plant growth, has not
had the attention of researchers in the same proportion as other
nutrients. This situation is presently changing due to several
reasons, namely: large increases in crop yields with a consequent
increased demand for this elements an increase in the use of S-free
fertilizers: decrease in the use of pesticides containing S: and the
use of energy sources with a relatively low S content, resulting in
lower S additions from the air.

This same situation holds true for Argentina, where some reports
of S deficiencies have appeared. These isolated reports will probably
begin to extend to greater areas as more studies are performed, and as
water, nitrogen, and phosphorus nutritional aspects are improved; thus,
S and perhaps certain micronutrients may become limiting factors for
plant growth. This problem could be particularly serious in farming
areas far away from industrial centers in which atmospheric-S inputs
would not be enough to compensate for soil-S deficiencies.

The main objectives of this study were: 1) to select and modify
methods for S analysis applicable to research on S needs in Argentina;
2) to compare the quantity of S extracted with water to that extracted
in a P solution; and 3) to study the effect of sample incubation upon

extractable S.

LITERATURE REVIEW

A large number of different methods to evaluate the sulfur status
in plants and soils are available from the literature of the last 20
years. These may be arbitrarily divided into methods of extraction
and determination as follows:
A. Total soil S extraction
B. Total plant S extraction
C. Extraction of available S
D. Quantitative determination of
extracted S
The word arbitrarily was used because most research work embodies
both extraction and determination procedures. Nevertheless, in order
to obtain a clearer picture of this subject, the above division will

be kept through this literature review.

A. Total Soil S Extraction
Two general methods for extracting
total soil S exist:
1. Reduction methods,

2. Oxidation methods.

1. Reduction Methods
Within this group, the technique most
oftenly used is that of Johnson and Nishita (1952). It is basically

a digestion of the soil sample at 115 C with a reducing mixture

composed of hydriodic acid, formic acid and red phosphorus. Most of
the S present in the sample is reduced to H23. This gas is then
passed through a pyrogallol and NaHZPOu gas washing solution by a

constant N flow, to finally react with a solution of ZnOAc-NaOAc.

2
Both, the gas washing and the ZnOAc-NaOAc solutions, must be free of
traces of heavy metals, particularly Cu, due to the precipitation of
GUS (Johnson and Arkley, 1954).

A detailed study of the reduction and color reaction steps has
been done by Gustafsson (1960a, 1960b). This researcher has modified,
among other things, the composition of the reducing mixture, by
substituting hypophosphorus acid for red phosphorus.

Several researchers, Spencer and Freney (1960), Dean (1966),
Searle (1968), Roberts and Koehler (1968), Walker (1972), and walker
and Doornenbal (1972), have used this technique.

Johnson and Nishita (1952), proposed this same technique for the
analysis of water samples, soil extracts, and plant samples.

Since this method does not distinguish between organic and
inorganic S forms, Freney (1957) adapted it for so” analysis of water
samples and soil extracts. He prOposed the use of bacteriological
filters to clean the sample of organic and inorganic colloids. Finally
the SOu-S was separated by precipitation, as BaSO . This researcher
also showed that the drier the soil samples were, prior to the
extraction, the more was the SOu-S extracted by water.

The reduction of soil elemental S to H S has been employed by

2
Barrow (1968), who extracted this element with chloroform and reduced
it afterwards with Fe powder in acid solution.

Magnesium metal in acid solution has also been used as a reducing

agent for S by Steinbergs et a1. (1962). They stated that total soil S
values obtained by this technique were smaller than those obtained by
oxidation methods.

Carson et a1. (1972), based on Dean's (1966) modification of
Johnson and Nishita‘s (1952) method, used a reducing mixture composed
of hydriodic, formic, and hypophosphorus acid, avoided the use of the
gas washing solution, and trapped the evolved H23 in NaOH, thus

simplifying the reduction-distillation apparatus.

2. Oxidation Methods

Different researchers have reported

trying many different oxidation mixtures to oxidize all forms of S to

-2
302, 303, and son .

Sodium peroxide has been used by Steinbergs (1955). Butters and
Chenery (1959) tried 111103 and Mg(N03)2 at 300 c for 10 hours. Bardsley

and Lancaster (1960) and Spencer and Freney (1960), used NaHCO3 at

500 c for 3 hours, and Steinbergs et a1. (1962) used Nch03 and AgZO at
550 c. Saalbach, Kessen, and Judel (1962) used KC103, and Bloomfield

(1962) proposed the use of V in a N flow; the complete oxidation of

205 2
the different volatile compounds was achieved by passing the vapors
through hot CuO. Chaudry and Cornfield (1966) used a HNOB-KNO3 mixture
at 550 C for 3 hours. Jenkinson (1968) used a three steps method to

evaluate total soil S. All S forms were first oxidized with KZCrZO7

and dehydrated H P04. The SO“ produced was then reduced to SO by

2
3P04. This gas was trapped in H20

3
heating with activated charcoal in H 2
where it formed H2804. The quantification of the acid was done by a

titrimetric technique. According to the author the overall sensitivity

of the method could be increased by means of a colorimetric instead of

the titrimetric determination. One of the drawbacks of this method was

that the hot H POu attacked the glass of the digestion flask, making it

3
necessary to replace the flask each 20 to 50 digestions.

The often used high-frequency induction furnace technique is also
an oxidation method. Searle (1971), Cadahia (1971), and Tiederman and
Anderson (1971), among others, have used this technique. Basically,
the soil sample is mixed with a catalytic oxidant mixture composed of

powdered Fe, M00 and CrO and heated by a high-frequency electrical

3 3'
flux induced in the sample itself, in which Fe acts as the conductor.

An 0 stream carries the S oxides produced, which react with a solution

2
of NaOH and are then measured by the Johnson and Nishita (1952)
technique.

Another method for total S determination is the one proposed by
Tabatabai and Bremner (1970a). The soil sampletndergoes a wet oxidation
with NaBrO and the oxidized S is quantified afterwards by the Johnson
and Nishita (1952) method.

Tabatabai and Bremner (1970b) also published a critical comparison

of several of these oxidation methods.

B. Total Plant S Extraction
Almost all the methods mentioned above
can be used for S analysis in plants, although some minor modifications
may be needed.
Among those not mentioned for soils, Lysyj and Zarembo (1958),
developed a technique to measure the S content of organic compounds,
adapted afterwards by Iismaa (1959) for plant analysis. This method is

based on the combustion of the sample in an Erlenmeyer flask with NaOH

and H 02: the S is then measured by Johnson and Nishita's (1952)

2
method.

The wet digestion with HNO3 and HClOu has been proposed by Shaw
(1959), the SOu-S being measured by gravimetry or any other available

technique.

C. Extraction of Available S
Several studies to correlate
extractable S with plant yields or S content in plants have been done.
Ensminger and Freney (1966) published a bibliographic review
(73 references) related to the diagnosis of S deficiencies by plant and
soil analysis.

The expression "extractable S" includes many S fractions. Williams
and Steinbergs (1958) correlated eight of these fractions (sou-s, total
water-soluble S, heat-soluble S, SO,+ soluble after ignition, alkali
soluble s, reducible S, and so“ released by H202 oxidation), with S
uptake and yield of dry matter of oats.

Spencer and Freney (1960) correlated S uptake of Phalaris tuberosa
with the following S fractions: acetate-extractable S, cold-water
extractable S, POu-extractable S, Aspergillus S, heat-soluble S,
hot-water-extractable S, reducible S, reserve S, and total S.

Good correlations were found between S uptake by plants and SOu-S,
total water-soluble S, and heat-soluble S in the research by Williams
and Steinbergs (1958): while Spencer and Freney (1960) found good
correlations with Aspergillus S and Pou-extractable S.

Fox, Olson and Rhoades (l96h) found that water-extractable S and

a 100 ppm P solution as Ca(H2POu)2 gave the bestagreement with sulfur A

values and total S uptake by alfalfa. However, they pointed out that
while both extractants behaved the same in unweathered soils, this did
not hold true for weathered soils (with sorbed SO“) in which the P
extraction correlated better with S uptake than did water-extractable S.
Magnesium chloride and LiCl were used by Roberts and Koehler (1968)
and CaCl2 by Walker and Doornenbal (1972) -- all found good correlations
between extractable S with these solutions and plant yield.
In a recent work, Hoeft, Walsh, and Keeney (1973) used four
for S extraction. The best

different P solutions, CaCl and Bray P

2 1
correlation with alfalfa yield was given by the S extracted by a
500 ppm P-2N HOAc solution, with the P as Ca(H2P0u)2. They also used a
regression equation to predict the yield response of alfalfa with good
results.

From this review it can be concluded that there is a tendency to
use salt solutions, particularly of P, to study correlations between
extractable S and plant yields. Total S and other artificial S fractions

have been set aside, while in some cases microbiological assays and

water-extractable S could still be used with some advantage.

D. Quantitative Determination of Extracted S
A great variety of
methods are actually available, even though none of them satisfy all
of the requirements of a good analytic technique, namely: accuracy,
precision, sensitivity, quickness and economy.
Although much has been done in the last 20 years in terms of
methods and instrumentation, S still offers some difficulties in its

quantification, especially with minute quantities.

The methods to be presented will be grouped according to the
physical, physical-chemical or chemical principle upon which they are

based.

1. Turbidimetry

 

One of the first applications of turbidimetric
methods to soils was that of Chesnin and Yien (1950), These authors
briefly reviewed an early application of it, particularly in the field
of medicine. To extract the soil S, they used the "Morgan solution"
(1941) formed by NaOAc and HOAc. Then, they precipitated the extracted

SO“ with 30-60 mesh BaCl crystals. The BaSO“ precipitate was

2
stabilized by a gum acacia suspension and the readings made in a
calorimeter with a blue filter. Reported results with this technique
were compared with those obtained by gravimetry.

Steinbergs (1953), found that the sensibility and reproducibility
of the method was somewhat affected by insoluble impurities present in

the BaCl used as the precipitating agent. These impurities acted as

2
crystallization nuclei causing a more homogeneous BaSO“ precipitate.

Due to the high purity of commercial BaCl Steinbergs (1955)

2.
prepared standardized solutions seeded with BaSO“ crystals and with
excess Ba. where these crystals acted as crystallization nuclei. He
added the unknowns or the standards to these solutions and read the
transmission 60 to 90 minutes after mixing. This same technique was
used to study different fractions of S as indexes of available S
(Williams and Steinbergs, 1959). According to the indicated procedure
no gum acacia solution was needed with this technique.

Since Mergan's solutions extracted colloidal organic matter which

interferred with the analytical method, Hesse (1957) proposed a
modification of Chesnin and Yien's (1950) method. The organic matter
was removed by precipitation with Fe(OH)3 and separated afterwards by
filtration. According to Bartlett and Meller (1960), when soil
extracts from Morgan's solution contained less than 20 ppm SO“

(6.7 ppm S), more than 6 g of BaCl2 per 50 m1 of extract should be
added to obtain a quantitative precipitation of the BaSO“.

A very detailed study of the best conditions to use turbidimetry
for total S in plants and soils has been done by Butters and Chenery
(1959). They studied the effect of such variables as size of the BaCl2
crystals, buffered medium to obtain a good dissolution rate of these
crystals, the use of H3P0h’ time of precipitation, and interferences
from other ions present. They also used the gum acacia solution as a
stabilizer, prepared according to Chesnin and Yien's (1950) procedure.

Massoumi and Cornfield (1963), used this method to analyze soil
extracts by introducing some modifications to increase its sensitivity.
With these modifications they were able to detect 2 ppm of S, (also the
method could be used for either total soil or plant 3).

Tabatabai and Bremner (1970a) compared turbidimetry with the
Johnson and Nishita (1952) procedure with good correlation.

Due to the need of a very careful standardization of this
technique in each of its steps plus the problem of the size of the

BaCl crystals, Basson and Bohmer (1972) proposed using the Technicon

2
autoanalizer for the SO“ turbidimetry. With this type of analyzer each
standard and each sample are treated in the same way, maintaining these
conditions through all the analysis. The authors used this method to

analyze sou content in plant samples digested with HNO3 and H0104.

10

Several researcher, for example, Bardsley and Lancaster (1960),
Dodgson (1961), Chaudry and Cornfield (1966), and Souty and Guennelon
(1967), have used turbidimetry as a mean for SO“ determination. Only
a portion of those who have improved this technique have been mentioned

here.

2. Nephelometry
An important contribution to the standarization
of this technique for SO“ determination as BaSOu was done by Toennies
and Balsay (1953). These researchers studied the effect of organic
solvents, pH and interferences of other ions on the sensitivity and
reproducibility with this method. According to the authors, the main
reasons for erratic results were: deficient washing of the tubes, SO“

contamination of the distilled water, and aging of the BaCl reagent.

2
Another modification of this method was by the use of organic
precipitants to precipitate the SO”. Based on early works by Belcher,

Nutten, and Stephen (1953), Belcher et a1. (1956), and Belcher and
Stephen (1959) -- who found that the u-amine uv-chlorodiphenyl chloride
(CAD) gave a highly insoluble salt with so” -- Martin and Stephen
(1967a) applied this reaction to the nephelometry of this anion. They
described in detail a technique to measure between 2.5 and 25 ppm of
SO“ (0.83 to 8.3 ppm of S) using CAD and a stabilizing solution of gum
ghatti. They also studied some possible anionic interferences. In a
short communication, Martin and Stephen (1967b) reported increased
sensitivity to 1 ppm detectable SO“ through improvement of the
stabilizing solution.

Stephen (1970) described the precipitation of SO“ with 2-amino

pyrimidine. He found this salt to be even more insoluble than CAD-Sou

11

forming regularly shaped and sized crystals with a very low tendency
to coagulate even after standing for several hours. The author
reported that SOLL concentrations as low as 0.05 ppm yielded a visible

precipitate.

3. Absorption Spectrometry
(As mentioned early, one of the most
widely used methods for soil and plant S analysis, is that of Johnson
and Nishita (1952). The H23 produced during the reduction step reacts
with the ZnOAc-NaOAc, yielding ZnS. Due to the acidification with a

sulfuric solution of p-amine dimetylaniline, the H S is liberated

2
reacting with the base (the resulting solution is colorless in the
reduced form, Bardsley and Lancaster, 1965). By adding Fe+3 ions, the
base changes to the oxidized form, generating a blue color (Johnson and
Urlich, 1959), which is measured with a spectrophotometer.

A detailed study of the experimental conditions that affect this
reaction has been done by Gustafsson (1960b).

In a recent work, Kirsten and Patel (1971) proposed the addition
of pyridine or sodium dodecyl benzene sulfonate to reduce or eliminate
the association among methylene blue molecules caused by S concentrations
exceeding 50 micrograms/100 m1 (Johnson and Nishita, 1952).

To increase the sensitivity of the method, they also proposed the
extraction of the methylene blue with a small volume of organic solvent,
namely 1,l,2,2,-tetrachloroethane and n-octanol-Z. In this way they
reported detecting as little as 0.2 micrograms of S.

Dean (1966) and Carson et al. (1972), collected the H S in NaOH.

2

Following this, a reagent composed by Bi(NO HOAc and gelatine was

3)3"
added, and the absorbency of the resulting solution measured with a

spectroPhotometer.

Bloomfield (1962), utilized p-rosaniline and a formaldehyde
solution to measure total S in soils colorimetrically.

Another technique is based on the reaction of S"2 with organic
disulfides. Several of these disulfides were studied by Humphrey
et a1. (1971) and according to the authors, between 0.05 and 21 ppm of

S can be detected with this technique.

a. Atomic Absorption Spectrometry
This method is generally based

on the precipitation of the SO“ ions with BaCl filtering the

2.
precipitate and measuring the excess Ba left in solution. Thus, the
amount of SO“ in the sample is determined by difference, (Warley and
Poou Yien Chin, 1970: Borden and McCormick, 1970: and Burge, 1970),

The major difference between the two first studies is that, in the
first one they did the conductimetric measurement of the extract to
approximate the concentration of SO , In this way they were able to
save time by taking aliquots that would fall within the range of the
standard curve: thus they did not need to make any dilution after

beginning with a group of samples.

5. Emission Spectrometry
This technique has been recently applied
by Forbes (1973) to measure Son in an indirect way (by the same
technique as in atomic absorption). adding excess BaCl2 to the sample
containing 304 with 50 % propan-Z-ol. Following precipitation and
centrifugation, the unreacted Ba content of the solution is measured by

its emission in a nitrous oxide-acetylene flame. According to the

13

author, SOu contents ranging from 0.5 to 5,0 and 1.0 to 10 ppm have

been measured with this procedure.

6. Potentiometric Titrations
This technique is based on the
titration of the sample containing son with a Pb salt solution, where a
change in potential indicates the end point of the titration.
Cortellessa and Napoli (1969) titrated the SO“ solution with a

Pb(N0 solution, after addition of potassium hexacyanoferrate II and

3)2
potassium hexacyanoferrate III, with a Pt wire as the measuring
electrode and a Ag reference electrode. At the equivalence point, the
precipitation of hexacyanoferrate II ocurred with a resulting sharp
change in the potential of the Pt electrode. Ross and Frant (1969)
and Goertzen and Oster (1972) used a Pb selective electrode and
titrated the son solution with a standard PbClOu solution. According

to the latter, 0,5 meq/l of SO“ could be measured with this technique.

7. Polarography

 

Khan and Webster (1968) based their technique on
the precipitation of Son in soil extracts as PbSOu. The excess of Pb
was measured by a drOpping Hg electrode polarized at a potential
corresponding to the diffusion current of Pb ions. According to the
authors they were able to measure a minimum Sou concentration of about
3 ppm, with a non-commercial polarograph.

Garber and Wilson (1972) applied differential polarography to

measure SO2 by bubbling the gas in a dimethyl sulfoxide solution.

19

8. X-rgy Emission Spectrggraphy
Robert and Koehler (1968a) used

this technique to analyze for S extractable with MgCl2 from soils of
Washington. They also did a very detailed study to determine its
application to soil S and compared it with the methylene blue method
(Roberts and Koehler, 1968b). According to the authors, this technique
makes possible the analysis of #0 soil extracts per day with a S
content varying between 0.5 and 14 ppm and a precision of :2 to £10 %
in the determinations.

Souty and Guennelon (1967), used the same technique to analyze the
3 content in plants, avoiding in this way the calcination of the
samples. They compared it with the oxygen bomb method followed by the
turbidimetric determination of the SO“ ions. Results by both methods
were comparable in samples containing 0.3 % S or more. In samples with
less S content, the oxygen bomb-turbidimetric determination was found
to be more reliable than the X-ray emission spectrography.

Another comparative study with four other methods has been done by
Tabatabai and Bremner (1970b) who used it for total soil S. They
reported that results with this technique were comparable with other
methods, but it had the disadvantage of a very time consuming

preparation of the sample if only S was to be analyzed.

9. Coulometry

 

Coulometric methods of analysis are based on the
exact measurement of the quantity of electricity that passes through a
solution during the occurrence of an electrochemical reaction. The
substance of interest may be oxidized or reduced at one of the

electrodes (primary coulometric analysis) or may react quantitavely in

15

solution with a single product of electrolysis (secondary coulometric
analysis), (Willard, Merrit, and Dean, 1965).

In a recent work, Bailey and Bishop (1972) applied the secondary
coulometric analysis to measure SO2 concentration in water. They used
a potentiometric method to find the end point of the redox reaction
that took place in solution, obtaining a 900 mV potential change at

this point, independent of the SO concentration. The sensitivity of

L.

2

the method lay within 10‘ to 10'5 molar, with a precision better than

1 %.

10. Volumetry
There are several published works that use this

principle to measure S in water extracts.

Belcher et a1. (1953) studied the possibility of using substituted
benzidines as precipitating agents for SO“. Later, Belcher and Stephen
(1959) developed a turbidimetric method using u-aminoh-chlorodiphenyl
(CAD) to precipitate sou which proved to be more insoluble than BaSO“.
In another work, Belcher et a1. (1956) described a technique by which
this anion was precipitated with excess CAD, and the excess reagent
titrated with NaOH using a mixed indicator to determine visually the
end point of the titration.

Fritz and Freeland (195a) also used volumetry for sou analysis in
aqueous extracts by a direct titration of the anion with a Ba standard
solution using alizarin red as the indicator. The titration with Ba has
been used by Armenau and Dragusin (1969) and by Ignatova and Abaturova
(1970), employing carboxiazo III or carboxisulfonazo III and nitrocromazo
respectively.

Sulfonazo III has been used as an indicator by Jenkinson (1968) in

16

the titration of the excess of added Ba with a standard K2304 solution.
When measuring elemental S in soils by reduction with Fe powder

in acid medium (Barrow, 1969), the H S produced was titrated with a

2
standard HgCl2 solution, using ditizone as an indicator.

Searle (1971). Cadahia (1971), Tiedeman and Anderson (1971), and
Tabatabai and Bremner (1970b), who worked with the oxidation of soil
and plant samples in a high-frequency induction furnace, also used

volumetry. The SO2 produced during the oxidation was trapped in HCl

and titrated automatically with a KI standard solution.

MATERIALS AND METHODS

A. Materials

1-§9_fls
During the present research three surface soils from
Michigan belonging to the series Morley, Hoytville and Miami were used,
Some of their characteristics are given in Table 1. These soils
developed under different drainage conditions. They present
differences in texture and organic matter content, but the pH among

them is very similar.

B. Methods

1. Preparation of the Samples
The soil samples were initially
air dried and then ground to pass a 10 mesh sieve (2 mm).

Two hundred grams of each soil, replicated six times, were
weighed and placed in plastic bags with hermetic sealing. Three
replications were brought to field capacity with S-free water and used
as controls. The rest were brought to field capacity with a K2804
solution, adding 0.100 g of S per 1000 g of soil (100 ppm) to each.
Both, the controls and the treated samples, were incubated for 0, 2,
and 6 weeks at 20 C. After incubation the samples were dried at 80 C

and ground to pass a 35 mesh sieve (0.5 mm).

17

18

.mpamao>flcb oveem semanoaz .Hnma .mamose .m.z .somaam shepozoocsz 509% spam AHV

 

 

er .N as we .0 a shoe eofiewe S or Hana:
m©.s mam om.w < heao decades zanoom oaaasehoz
mo.m mom wm.m < smoa tho vocasnd Ham: hoHHoz
e as
Hopes: owmcasnm
oasswno gm mm couanom oHSpXoe Handpez moanom Hfiom

 

 

Adv .eoms madam semasoaz m on» mo moapuomonm Hmoamhnn use Hmoasono .moahom Haom .H manna

19

2, Extraction Procedures

a. Water extractable S

Reagents

Sulfur-free water: pass distilled water

 

through different resins to extract organic matter, cations, and
anions, (organic removal, D8904: cation removal, D8905: and mixed bed,
D8907, from the Barnsted Sybron Corp., Boston, Mass.), and finally
through an anion exchange resin in the Cl form, (Amberlite IRA-400 Ar).
Procedure

Weigh 20 g of sample and place in 100 ml
centrifuge tube: add 50 m1 of S-free water (soil:water ratio 1:2.5) and
shake for #5 minutes on a wrist action shaker. Afterwards, centrifuge
for 30 minutes at 2200 rpm, (an International centrifuge, Mbdel 5V was
used) and then filtrate through Whatman No. #2 filter paper. Each soil

sample was extracted in triplicate,

b. Extractable S with a 500 ppm P solution
Reagents
Extractant solution: prepare a 500 ppm P

solution in 2 N HOAc by dissolving 2.03u g of Ca(H2POu)2.H20 in
114.38 ml of HOAc (glacial) in a 1 liter volumetric flask. Dilute to
1 Liter with S-free water.

Procedure

The same procedure outlined for water

extractable S was used for these extracts.

20

c. Water extractable SO“

 

Reagents
1 N sodium hydroxide: dissolve 40 g of NaOH

in S-free water and make to 1 liter.
5 % ferric chloride solution: dissolve 5 g

of FeCl .6H20 in 100 ml of S-free water.

3
Procedure
While shaking add 2 ml of FeCl3 solution to

25 ml aliquot of the extracts obtained according to 2.a. Then add
0.5 m1 of the l N NaOH, shake and let the suspension stand for 2 or 3
minutes. If the supernatant liquid is still turbid and/or colored, add
one drop of the base, shake and let stand for another 2 or 3 minutes.
Repeat this procedure until the supernatant liquid is perfectly clear
and colorless. Then filter the solution through Whatman No. 41 filter
paper.

The pH of 54 extracts clarified according to the indicated procedure
in every case was between 6 and 6.5.

The present technique is based on Hesse's (1957) work, and was

adapted to use with aqueous extracts.

d. Extractable son with a 5004ppm P solution

 

Reagents

Washed activated charcoal: wash approximately
20 g of activated charcoal with 500 ml of 0.1 N HCl. Filter under
vacuum with a Buchner funnel.
Procedure
Measure 25 m1 of the extracts obtained

according to 2.b. and place in a small Erlenmeyer. Add a small amount

21

of washed activated charcoal, shake and filter through Whatman No. 42
filter paper. Controls using S-free water following this procedure

did not show detectable SO,4 levels.

3. Total S in Soils
Tabatabai and Bremner's (1970a) procedure

was followed for the alkaline oxidation step, and Gustafsson's
(1960a, 1960b) for the reduction and color reaction steps. As these
two last steps are common to several types of extraction procedures
for both, total and extractable S, they are describaias an independent
method in Section 4.

Modifications introduced to these procedures are justified in the

next chapter.

Alkaline oxidation (Tabatabai and Bremner, 1970a)

 

Apparatus: the digestion flask corresponds to Tabatabai
and Brenner's (1970a) distillation unit.
Reagents
2 N sodium hydroxide: dissolve 80 g of NaOH in
S-free water and make to 1 liter.

Sodium hypobromite solution: add slowly with

 

continuous stirring (approximately 0.5 ml/min) 3 m1 of Br2 to 100 m1 of
2 N NaOH, This reagent should be prepared fresh before using it.

Formic acid: concentrated, analytical reagent

 

grade,
Procedure
Place sufficient sample to have between 10 and

60 micrograms of S (approximately 0.1 to 0.3 g of soil) in a digestion

22

flask. Add 3 ml of the NaBrO solution, mix the contents gently and let
stand for 5 minutes. Mix the contents again and place the flask in a
sand bath at 250-260 C. Once the residue is dry (15 to 20 min), leave
it in the bath for 30 additional minutes at the same temperature. Then
remove the flask from the bath and allow it to cool at room temperature:
add 1 m1 of formic acid and continue with the reduction step (see 4.a.).
Perform a blank for each set of samples.
4, Sulfur Determination_by the Methylene Blue Method
(Gustafsson, 1960a, 1960b).
a. Total S determination
Appaaatus: A diagram of the reduction-distillation
unit is shown in Tabatabai and Bremner's (1970a) work.
Reagents: For a complete list of reagents see
Gustafsson (1960a, 1960b).
Procedure
Place 15-20 m1 of S-free water in the gas

washing column, wet both ends of the U tube with a drop of H P0“

3
(Tabatabai and Bremner, 1970a) and attach it to the unit. Immerse the
gas outlet tube in a 100 ml volumetric flask -- containing 65 ml of
water and 10 ml of the ZnOAc-NaOAc solution -- with the end of the
tube at l or 2 mm from the bottom of the flask. Begin the water

circulation in the condenser and the N flow, regulating it to obtain

2
150-200 ml/min with 0.75 atm of pressure.

Wet the ground glass connection of the unit with one drop of
HBPQ4' Add 4 ml of the reducing mixture to the flask that contains the
digested sample with formic acid, and connect it to the unit by means

of the two springs.

23

To prevent overheating, place an asbestos board fitted with a 3 cm
diameter hole in the middle in contact with the flask, then light the
burner regulating the tip of the blue flame to finish a few millimiters
from the bottom of the flask. The sample should boil within 60 seconds
and the distillation should be continued for 10 to 15 minutes.

Once this period is finished, disconnect the volumetric flask,
leaving the gas outlet tube in the flask with its latex connection.
Then add 10 ml of the methylene blue solution through the tube with a
quick delivery pipette forming a layer in the bottom of the flask:
raise the tube and rinse it with one or two jets of S-free water.
Rotate the flask gently and immediately add 2 ml of the Fe(NHL,)(S0,,)2
solution with a quick delivery pipette. Close the flask rapidly and
shake it vigorously for 30 seconds. Then bring to volume with S-free
water and leave it 15 minutes or more in subdued light for color
development.

It is important to note that the time between the addition of the
first solution and the shaking of the flask should be as short as
possible to prevent gas losses.

All the readings were done with a Bausch and Lomb spectrophotometer
(Model Spectronic 20) at 667 nm. Results were compared with those of a
standard curve with 20, 40, 50, and 60 micrograms of S.

The unit was cleaned by attaching the gas outlet to the distillation

unit and applying vacuum in the N inlet, immersing the gas outlet in a

2
beaker with S-free water 3 or 4 times. Once finished, the gas washing
column was drained of the remaining water and refilled to the 15-30 ml

mark by carefully applying vacuum. Once the digestion flask was

disconnected, the unit was ready for another reduction.

24

Generally some water remained in the N2 inlet tube and between
this tube and the ground glass connection. This would cause dilution
of the reducing mixture, therefore, the N2 inlet tube was washed with
one or two jets of acetone and the inner side of the ground glass

connection dried with absorbing paper,

b. Extractable S determination
Place an aliquot
containing between 10 and 60 micrograms of S in the reduction flask.
Dry the content in an oven at 120-130 C. Cool the flask to room
temperature and add 4 m1 of the reducing mixture. Then follow the same

procedure as described above.

5. Sulfate Determination by I Colorimetry

2

Reagents
Potassium iodide: KI crystals of reagent quality.

Acetonitrile: glass-distilled quality.

 

Barium iodate: dissolve 10 g of KIO3 in 500 ml

 

of S-free water and 5 g of Ba(N03)2 in another 500 ml of S-free water.
Mix both solutions and stir for 4 or 5 minutes. Let it stand and
eliminate excess water. Repeat these washings with water 3 or 4 times.
Do the last one with ethyl alcohol, filter under vacuum and dry the
residue in an oven at 90 C.

Sulfur standard solution: dissolve 5.43 g of
KZSOu (dried at 105 c for 2 hours) in S-free water and dilute to

1 liter. This solution contains 1000 ppm S as SO“,

25

Procedure
Place 16 m1 of acetonitrile in a 50 ml
Erlenmeyer, add an aliquot containing between 3 and 15 micrograms of
SOn-S and dilute to 20 ml with S-free water. Add about 30 to 40 mg/of
Ba(103)2 to the solution and let it stand for 1 hour, shaking
occasionally.

Filter afterwards through Whatman No, 42 filter paper, receiving
the filtrate in a 25 ml volumetric flask. Add some KI crystals, shake
until dissolved, and then add 2 m1 of glacial HOAc.

Dilute to the mark with S-free water and measure after 10 minutes
at 400 nm. Perform a blank for each set of samples.

Compare the results with those of a standard curve with 3, 5, 10,

and 15 micrograms of 804-8 using the same procedure.

RESULTS AND DISCUSSION

1. Total S Determination
In the present investigation,
Tabatabai and Bremner's (1970a) method was used.

Two modifications were introduced to their original procedure in
an attempt to reduce the time of analysis per sample and to increase,
in some cases, the amount of S measured.

Tabatabai and Bremner's (1970a) original contribution was in the
procedure to oxidize the organic and inorganic S present in the sample
and also in the modification of the distillation unit. In the second
part of the method -- SO“ reduction and color reaction steps -- they
used the method of Johnson and Nishita (1952), in which these authors
proposed a one hour reduction-distillation period. Nevertheless,
Gustafsson (1960a) in a very detailed study of the reduction step,
suggested that a 10 minute reduction-distillation period was enough

for the S reduction and transfer of the H S formed.

2
Since this would mean a great time saving per sample, some
preliminary studies were performed to determine if accuracy was still

mantained. Results can be seen in Table 2.

The analysis of variance (Table 3) shows that the recovery of S
between 10 minutes and 1 hour of reduction-distillation did not vary
significantly, at the l.% level, for any of the soils. Therefore,

a 10 minute reduction-distillation was adopted for all the analysis

performed by this method.

26

27

m.a m.sH era
m.s m.sH boa
s.m m.m rma

hmv>o mm seas
concedes

.wopsoaamon 03» mo omeo>w .Hdom mo spam Hon m mo mssnmoHOfiz A*v

s.m m.s ram
m.m m.m~ men
is «.3 in

eaves mm seam
nonsense oz

o.~ m.s mam
H.N a.a mas
o.: m.s sea

33% mm seas
soapsaan

m.s m.sa mmm
s.a r.mH was

b.H s.m ANN

Amuse mm mesa
cosssaae oz

 

 

user H

mopssas 0H

 

cane soasoaaaenaa-soaeoseom

‘

seas:
Easier

aoHHoz

moewom HHom

 

 

.coaemeaswosoe m anon

Hoses coca sosssaae essence mo ens soasnfiaacnaeusoaposeow mo been or» no mososamsH .N canoe

Table 3. Analysis of variance of experiment of the influence of the
reduction-distillation time and of residue dilution.

 

 

 

 

 

 

Source of Degrees of Sum of Mean F

Variation Squares Squares

MORLEY SERIES

Treatments 7350.0
Time 13.9 13.9 0.15
Dilution 7306.2 7306.2 82.16 *
Interaction (TxD) 29.9 29.9 0.33

Experimental error 355.7 88.9

Total 7706.3

HOYTVILIE SERIES

Treatments 4098.0
Time 680.2 680.2 2.10
Dilution 7 o 3 7 o 3 O. 02
Interaction (TxD) 3410.5 3410.5 10.52

Experimental error 1259.0 323.9

Total 5393.0

MIAMI SERIES

Treatments 1971.0
Time 749.1 749.1 4.67
Dilution 1174.4 1174.4 7.4?
Interaction (TxD) 47.5 47.5 0.30

Experimental error 628.? 157.2

Total 2599.7

 

(*) Significant at the l % level.

29

The second modification is related to the dilution of the sample
as proposed by Tabatabai and Bremner (1970a).

According to their procedure, once the alkaline oxidation is
finished, and with the flask at room temperature, 1 ml of S-free water
should be added to the residue heating for a few seconds in order to
resuspend it. After this, 1 ml of formic acid and 4 ml of the
reducing mixture are added and the reduction-distillation begins
immediately.

In the work by Gustafsson (1970a), already mentioned, the author
stresses the effect of diluting the reducing mixture. She found that
the addition of 0.5 m1 of water to an evaporated SO” sample decreased
the recovery of son by 6 %. Two mililiters of water diminished the
yield ' by ‘35 96.

Results from preliminary tests conducted during this investigation
were not so conclussive. It can be seen in Table 3 that only one soil
(Morley series) showed a significant difference at a l.% level not
favoring the dilution of the sample.

Nevertheless, from Table 2 it is clear that the sample of the
Miami series shows the same tendency, although at a different level of
significance, (6 %).

The most difficult results to explain are those of the Hoytville
series which showed a significant interaction between the two variables
at the 5 % level, thus not allowing a separate analysis of the behavior
of both variables.

Therefore, although the results of the analysis of variance at
the l.% level of significance would allow the modification of the

original procedure with respect to the dilution of the oxidation

3O

residue, they are not sufficiently conclusive. It might be suggested,
thus, that more soils should be analyzed with respect to this behavior
before adopting a definite procedure.

A side effect of this latter modification was seen during the
washing of the digestion flasks. The formic acid did not suspend the
residue nor did the reducing mixture, therefore leading to difficulties
in washing the glassware.

This problem, although not sufficiently tested, was overcome by
reversing the order of addition of the reagents, namely adding first
the reducing mixture, then the formic acid, and beginning the
distillation immediately. No variation in the results was observed
when this procedure was followed.

The values of total S for the three soils analyzed can be seen in
Table 4. The percentage of recovery for the samples which received
100 ppm of SOQ-S are also given in this table.

These values are sufficiently accurate and comparable with those
given by Tabatabai and Bremner (1970a), if the differences in the S
application techniques are considered. These authors added the organic
and inorganic S to the subsample to be analyzed (0.1 to 0.2 gr of soil)
with 1 m1 of water, while in the present investigation the S was added
to 200 gr of Soil, of which the subsamples for analysis were withdrawn
after a 2 weeks incubation period.

Considering the sources of error -- sampling, subsampling,
weighing error, intrinsical error of the method, heterogeneity due to
an uneven mixing of the added S, instrumental error and reading error--
the variation coefficients are within acceptable limits, (except for a

6.6 % value, all the rest are below 4 %). With a better standarization

31

.mopsoaaaoh eons» mo owwho>< A*V

 

 

m.m m.mOH 0.0H OOH son OOH HseHz
O.N N.OOH H.mH OOH smm OOH oHHHeegor
O.H OOH O.m OOH mam OOH sono:
0.0 --- 0.0H -u- HON OON HseH:
r.H -u- s.s --- an: OON oHHHseaor
a.m -u- N.@ nu- «ma OON sonoz
m R Egg *EAHAH WE
. o. huh >0 . . ¥SWHO3
s O o com a m m Omens m Hence oHasmm nonom HHom

 

 

.m coeds mo hHo>oooH seconds use m HHom Hopoe .3 capes

32

of the technique these values could have been kept below 3 %. All
results were reproducible.

Based on the accquired experience, three points could be subject
of further standarization:

a. Nitrogen flow: Although disagreement

 

exists among researchers as to which is its Optimum value, the
important factor, within limits, is to have the same flow for each unit.
Thus, a good calibration at each outlet is required, particularly since
reducing the reduction-distillation time makes this variable more
critical,

b. Heatiag: According to Johnson and
Nishita's (1952) technique, an asbestos board is interposed between the
flame and the distillation flask. Gustafsson (1960a) fitted the board
with a 3 cm diameter hole in its middle part. With this modification
she obtained a faster boiling of the solution while preventing
overheating of the flask walls. This researcher also reported an
incomplete S reduction the to distillations carried without boiling the
solution,

Nevertheless, care should be taken not to allow direct contact
between the flame and the flask, thus, overheating the sample. During
the first distillations of standard S solutions, working with 4 units
simultaneously, bad flame regulations allowing overheating of the
flasks yielded losses in the S recovery.

Therefore, use of good quality burners is recommended, in which an

accurate regulation of the flame is permitted.

33

0. Temperature regulation during_the color

 

development reaction: Gustafsson (1960b) studied the influence of this
variable on the color reaction step. Her conclusions are summarized
as follows: A compromise relation exists between the temperature of

the solution before the addition of the reagents and the intensity of

the blue color formed. With increasing temperatures, the H S tends to

2
escape to the gaseous phase, therefore diminishing the color intensity.
With lower temperatures this problem is minimized but the speed of the
color formation reaction is much slower, thus allowing secondary
reactions to occur which reduce the yield of the methylene blue color.

According to the author, some researchers have found an Optimum
temperature for the reaction at 24 C, although she does not confirm
those same results.

Thesecond compromise relation is established between the
temperature at the moment of measurement and the absorbance obtained,
To measure this effect, Gustafsson (1960b) worked with a thermo-
regulated ceILcompartment in the spectrophotometer. Her conclusions
were that in this case the temperature effect was positive, that is,

the absorbance increased with the temperature. She proposed 20 C as the

optimum temperature to satisfy both relations.

2. Extractable S Determination

 

Sulfur-free water and a 500 ppm
P Solution were used as extractant solutions. Gustafsson's (1960a,
1960b) procedure was followed for the analysis of the extracted S. The
comments regarding this technique were already made and they will not

be discussed further: the three points mentioned —- N flow, heating,

2

34

and temperature regulation in the color reaction step, apply equally
here.

The results obtained from the analysis of the aqueous extracts and
of the P solution extracts can be seen in Tables 5 and 6, respectively,

The extracts were done in triplicate in all cases, and the reported
results are the average of each triplicate. This factor must be taken
into account, particularly when analyzing the standard deviations.

These are, in some cases, somewhat high due to the fact that they
include both variation "among extracts" and "within extracts". Therefore
confounded effects exist that could have been avoided by analyzing each
extract two or three times. In this way, variability would have been
divided into "within extracts" (errors of the method, instrumental, and
reading error), and "among extracts" (sampling, subsampling and weighing
error). This procedure was not followed due to a lack of time.

The variable studied was the change in extractable S with time of
incubation. The results obtained were analyzed in the following way:

a) Test of hypothesis using the F distribution of Snedecor that
relates two sample variances, in order to compare two p0pulation
variances, (Cappelletti, 1972), None of the variances compared in pairs
showed significant differences at a 5 % level and were considered
homogeneous from a statistical point of view.

b) Comparison of treatment means using the Student distribution
with the pooled estimate of the common variance for the treatments under
consideration. The comparisons were made between the means of samples
with and without incubation, and for both extractants.

Asterisks in Tables 5 and 6 show the soils in which there was a
significant increase, at the 5 % level, in the S content due to

incubation.

35

Table 5. Water-extractable S,

 

 

 

Soil Series Added S Incubation Extractable S S.D. C.V.
ppm weeks ppm (1) 9r

Morley 0 0 10.5 1.0 9.6
.. 0 6 16.0 * 1.2 7.4

" 100 O 96.5 8.4 8.7

.. 100 6 112.6 6.5 5.7
Hoytville 0 0 13.5 (2) 1.3 9.8
.. o 6 22.8 * 0.8 3.4

.. 100 0 101.4 7.4 7.4

.. 100 6 113.4 3.7 3.2
Miami 0 0 7,4 0.4 6.1
.. 0 6 13.5 * 0.8 6.3

.. 100 0 103.0 4.1 4.0

" 100 6 119.5 * 7.3 6.1

 

(1) Average of three replicates.
(2) Average of two replicates.

(*) Significant difference at the 5 % level between treatments.

Table 6.

36

Extractable S with a 500 ppm P Solution.

 

 

 

Soil Series Added S Incubation Extractable S S.D. C.V.
ppm weeks ppm (1) %

Morley O 0 12.6 0.6 5.2
" 0 6 21.3 * 0.9 4.1

" 100 0 114.0 7.0 6.2

" 100 6 112.5 6.8 6.1
Hoytville 0 0 14.4 0.7 4,9
.. 0 6 28.3 * 1.3 4.6

" 100 117.1 3.8 3.2

" 100 6 110.9 7.9 7.2
Miami 0 O 8.6 0.3 3.6
" 0 6 18.0 * 0.5 2.8

" 100 115.9 3.1 2.6

" 100 6 119.1 6.0 5.0

 

(1) Average of three replicates.

(*) Significant difference at the 5 % level between treatments.

37

All the samples without added S showed significant differences
with both extractants. Of those samples with 100 ppm S added, instead,
only one belonging to the Miami series showed a significant increase
when using water as an extractant.

Considering that the method used to measure S embodies both, the
organic and inorganic forms of this element, its increase with both
extractants after incubation could be explained by the decomposition of
the organic matter originally present in the soils.

Williams (1967) studied the effect of several factors upon the
organic S fraction during incubation. Two bibliographic reviews
related to this matter were done by Starkey (1966) and by Freney and
Stevenson (1966). From these studies, it can be concluded that the
soil organic S fractions suffer biodegradation processes, as happens
with N, P and other nutrients. These processes would lead to the
formation of simpler 8 compounds and of inorganic S, that would be
extracted more readily by several extractants.

With the method employed to measure S it is not possible to
determine which of these two forms is most important in explaining the
increase of this element after incubation.

Nevertheless, if the amounts of S extracted with water and P in
the control samples are compared, one may conclude that the
differences of 1 to 2 ppm of extracted S between both solutions is due
to the greater exchangeable power of the P solution as compared to the
water. This solution develops an acid pH at equilibrium, therefore
intensifying anionic exchange, in which the H2P0; tends to replace the
sorbed SO“ (Reisenauer et al. 1973).

On the other hand, the differences in water extractable S between

38

0 and 6 weeks of incubation were 5.5, 9.3, and 6.1 for the Morley,
Hoytville, and Miami series, respectively. In this same order, the
differences using the P solution were 8.7, 13.9 and 9.4 ppm.

Based on the prior discussion it is suggested that the differences
in water-extractable S are due to an increase in the amount of simpler
S compounds plus a fraction of the completely mineralized S. The
increase in P-extractable S could be explained by the fact that the
P solution extracted the same fraction as the water, plus that amount
of mineralized S subsequently sorbed.

The analysis of the results in those samples with 100 ppm of
added S becomes more difficult because it is not possible to
differentiate between the S originally present in the soil from that
added.

Nevertheless, it is worthwhile to compare the results of both
extractants with the unincubated samples: A sorption phenomena can be
noted, particularly in the Morley and Hoytville series, when water was
used as the extractant, which may be explained by their higher clay
content. This process was not as evident in the Miami series,
especially when the standard deviations are considered for this set of
values. The samples extracted with the P solution did notshow the
sorption process due to the greater exchange power of this extractant.
Therefore the S originally present plus the added, was recovered in the
unincubated samples.

This SO“ sorption process has been related to the presence of
organic matter, free Fe- and Al-oxides, (Chao, Harward and Fang, 1962):
in the acid range it was essentially proportional to the content of

hydrous Fe- or Al-oxides, although when both were present, Al—oxides

39

were more effective in this process (Chao et al. 1964). The Fe system
exhibited increasing adsorption with increased acidity whereas the Al
system Showed a maximum at pH 4.0. Dissolution of the hydrous Al-oxide
phase decreased the SO“ adsorption in Al-coated soils when going below
this pH, (Chao et a1. 1964). The nature of soil acidity was shown to
be important in that the Al-clays retained more SO4 than H-clays,

(Chao et a1. 1962), and according to these authors the possible
mechanisms of 504 adsorption would be given by:

a. Anion exchange due to positive charges developed on hydrous
Fe- or Al-oxides, or on the crystal edges of clays, especially
kaolinite, at low pH's.

b. Retention of SO“ ions by hydroxyl-Al complexes by coordination.

0. "Salt adsorption" resulting from attraction between the
surface of soil colloids and the salt.

d. Amphoteric properties of soil organic matter.

Chang and Thomas (1963) pr0posed a mechanism that involved a
cation and anion reaction: due to it, K+ would be adsorbed by exchange
and/or hydrolysis of Al3+ on the clay surfaces, giving H+ into
solution. At the same time, 80:- would replace OH- ions from R(OH)
coatings on the clay and substitute for them. The replaced OH- ions
will react with the H+ ions. Whether or not there is a change on the
system's pH would depend on the relative rates of the two reactions.
For the SO“ retentive soils, it has been shown (Fang et a1, 1962, Chao
et a1, 1962) that the 30h adsorption data can be fitted into either the
Freundlich or the Langmuir isotherms, depending on the range of SO”
concentrations of the equilibrium solutions.

After 6 weeks of incubation, this sorption process was apparently

4O

compensated by mineralization. The Hoytville series was the only one
that did not recover the S level of the control with incubation. The
increase in water extractable S after incubation in the Miami series
samples was Significant at the 5 % level, being the only case of all
the samples with added S.

The behavior of the P extractant varied drastically as compared to
water, in the samples with incubation. The original values were
recovered only in the Miami series, and no indication of a
mineralization process was shown by any of the samples. In fact, this
is notpossible because if mineralization occurred, as appears from the
extraction with water, it should have also been shown by the other
extractant solution.

The similarities in the values of S extracted with both extractants
in the samples with 6 weeks of incubation should be noted.

Somewhat high standard deviations and the small number of soils
tested did not allow a definitive conclusion with respect to this

aspect.

3. Extractable SO,+ Determination

 

 

As was mentioned before, several
techniques are available for the SO“ analysis -- the turbidimetric one
probably being the most widely used.

The indirect measurement of this anion using atomic absorption
spectrOphotometry is presently being developed.

The lower detection limit is the principal problem of both
techniques as neither one will measure concentration of lppm or less
with sufficient accuracy.

In a recent work, Hinze and Humphrey (1973) used a so called

41

amplification reaction, which consists of precipitating the 804 with

Ba(103) -- a relative insoluble reagent -- as BaSOLP (even more

2
insoluble). Two moles of IO

3

present. In an acid medium and with excess 1', 3 moles of 12 are

are produced for each mole of SO“

formed per each 10; present.

Other anions have been previously measured with this technique,
and Bertolacini and Barney (1957, 1958) used Ba chloranilate for the
precipitation of SO“. They measured the liberated chloranilate ion
with visible and ultraviolet spectrophotometry.

Hinze and Humphrey (1973), following Lamber and Zitomer's (1963)
procedure, used a reagent composed of starch and CdI2 to measure the
I formed in the reaction, The I

2 2
blue compound. Volumetry with thiosulfate would be another way to

reacts with the starch, forming a

measure the I2 formed (Weisz and Fritsche, 1970),

Detection limits are 1.2 and 0.5 ppm of SOLp (0,4 and 0.16 ppm S)
for the Hinze and Humphrey (1973) and Weisz and Fritsche (1970) methods
respectively.

The first method was tried during this investigation. It was a
highly sensitive method but offered some problems for its standarization.
These were, basically:

a. Somewhat high blank values, due to the solubility of the
Ba(103)2 in the alcoholic solution.

b. Turbidity of the starch-iodide reagent that would not disappear
upon dilution.

c. Difficulties in reproducing a standard curve.

It was thought, therefore, that the I could be measured by direct

2
spectrophotometry of its yellow color.

Several tests and experiments were conducted, and although no

42

definitive conclusions were obtained, the possibility exists of having
found a very simple and accurate method, of a high sensitivity for $04
and eventually also for 103. Nevertheless, it is important to point out
that the procedure outlined in the preceding chapter is by no means

a definitive one.

To improve the results, several variables were studied: Type of
solvent, optimum solvent pr0portion to minimize the blank values,
effect of evaporating the solvent in the absorbance values, optimum
dilutions, amount and type of acid to drop the pH. The advantage of
adding the K1 as crystals or as solution and the stability of the color
as a function of time were also studied.

A standard curve is shown in Figure 1, where each value is the
mean of triplicate standards. The molar absortivity calculated as a
function of S concentration was 21,950. In this curve, zero
absorbance corresponded to the blank. Its absorbance value, when
compared with an 80 % acetonitrile solution was always below 0,013
(greater than 98 % transmitance).

The minimum concentration used was 0.15 ppm SOu-S. A tendency to
lose linearity was observed when trying to go above 0.7 ppm SOu-S, so
other dilutions should be tried.

The continuous Spectrum between 460 and 320 nm can be seen in
Figure 2. It was made in a Beckman automatic spectrophotometer, Model
DB-G. Routine measurements were done with a Bausch and Lomb
spectrophotometer, Model Spectronic 20. The minimum transmission peak
(350 nm) could not be used for them, because the blank would notread
100 % transmitance at this wavelength; therefore, the 400 nm band was

used,

43

Absorbonce

0,50

0.40

0,30

0,20

0,l0

 

 

0,00

 

3 5 l0 l5
Microgroms 8/20 ml

Figure I. Relationship between absorbance and 8 concentration. ( x = 400 nm, d = l cm ).

Transmittonce (°/o)

IOO -

80L

60~

40

I

I

20

 

 

480 440 400 360 320
Wavelength (c m)

Figure 2: Absorption Spectra of the iodine/ acetonitrile system
(80% acetonitrile, 20 °/o water, 5 ug lOé/ 25 ml, d= lcm),

45

Two aspects should be further studied, namely:

a. Optimization of the conditions for the 804 precipitation as
BaSO“.

b. Possible interferences from ions commonly present in soil
extracts.

Although several extracts obtained according to the procedure
outlined in the preceding chapter were analyzed, the results will not
be presented until the points mentioned above have been studied in

detail.

SUMMARY AND CONCLUSIONS

Samples of three Michigan surface soils -- Morley, Hoytville, and
Miami series -- were incubated 0, 2, and 6 weeks at 20 C.

Half of the samples received 100 ppm of S (as K2804), and the
other half were used as controls.

Total S content was measured in the samples after 2 weeks of
incubation, with and without added S.

The samples with 0 and 6 weeks of incubation -- with and without
added S -- were analyzed for their content of extractable S, with water
and a 500 ppm P solution as extractants.

The following conclusions were drawn:

1. The total S content of
the Morley, Hoytville and Miami series were 222, 424, and 201 ppm of S,
respectively.

2, The percentage S recovery
in the samples with 100 ppm of added S was 100, 103.5, and 106.2,
respectively.

3. The results of preliminary
tests with the alkaline oxidation method allowed for the reduction of
the time of analysis per sample from 1 hour to 10 minutes.

4. The data suggested that
the dilution of the residue with water should not be done for all type
of samples. It was shown that in some cases this would cause a reduction

in the recovery of total S.

46

47

5. A significant increase in
the S content, at the 5 % level, due to 6 weeks of incubation was found

in all the samples, using both extractants.

6. In the samples with
100 ppm of added S, these same differences were significant only in the

Miami series samples where water was used as the extractant.

7. A new method for 804
analysis was developed. Although still in its standarization stage,
SOLL concentration of as little as 0.45 ppm was measured with good

reproducibility.

BIBLIOGRAPHY

BIBLIOGRAPHY

Armenau, V. and Dragusin, I. 1969a. Study of some azo-derivatives
complexes of chromotropic acid with Ba. I. Rev. Roum.Chimie

14:1367-1375.

1969b. Study of some azo-derivatives complexes of
chromotropic acid with Ba. II. Rev. Roum. Chimie 14:1511-1517.

Bailey, P.L. and Bishop, E. 1972, The coulometric determination of
sulfur dioxide using differential electrolytic potentiometry for
end-point location, Analyst 97:311-314.

Bardsley, C.E. and Lancaster, J.D. 1960. Determination of reserve
sulfur and soluble sulfates in soils. Soil Sci. Soc. Am. Proc.
24:265-268.

Barrow, N.J. 1968. Determination of elemental sulfur in soils.
J. Sci. Food Agric. 19:454-456.

Bartlett, F.D. and Neller, J.R. 1960, Turbidimetric determination of
sulfate-sulfur in soil extracts. Soil Sci. 90:201-204.

Basson. W.D. and Bohmer, R.G. 1972. An automated procedure for the
determination of sulfur in plant tissue. Analyst 97:266-270.

Belcher, R., Nutten, A.J., Parry, E. and Stephen, W.I. 1956. The
titrimetric determination of sulfate with 4-amino 4'-chloro
diphenyl hydrochloride as reagent. Analyst 81:4-8.

Belcher, R., Nutten, A.J., and Stephen, W.I. 1953. Part XII.
Reagents for the precipitation of sulfate. J. Chem. Soc. pp.
1334-1337.

Belcher, R. and Stephen, W.I. 1959. The use of 4-amino 4'-chloro
diphenyl as reagent for the determination of sulfates. Anal.
Chim. Acta 20:197-199.

Bertolacini, R.J. and Barney, J.E. 1957. Colorimetric determination
of sulfate with barium chloranilate. Anal. Chem, 29:281-283.

. 1958. Ultraviolet spectrophotometric determination of
sulfate, chloride, and fluoride with chloranilic acid. Anal.
Chem. 30:202-205.

Bloomfield, C. 1962. A colorimetric method for determining total
sulfur in soils. Analyst 87:586-589.

48

49

Borden, F.Y. and McCormick, L.H. 1970. An indirect method for the
measurement of sulfate by barium absorption spectrophotometry.
Soil Sci. Soc. Am. Proc. 34:705-706.

Butters, B. and Chenery, E.M. 1959. A rapid determination of total
sulfur in soils and plants. Analyst 84:239-245.

Cadahia, C. 1971. Semiauthomatic determination of organic and
mineral sulfur in agricultural samples with the Leco apparatus.
Anales de Edafologia y Agrobiologia 3O (7/8):8l5-826.

Cappelletti, c. 1972. Elementos de Estadistica, pp. 108-110.
Ed. Cesarini Hnos., Buenos Aires, Argentina.

Carson. J.A., Crepin, J.M. and Nemunis-Siugzdinis, P. 1972. A
sulfate-sulfur method used to delineate the sulfur status in
soils. Can. J. Soil Sci. 52:278-281.

Chang, M.L. and Thomas, C.W. 1963. A suggested mechanism for sulfate
adsorption by soils. Soil Sci. Soc. Am. Proc. 27:281-283.

Chao, T.T., Harward, M.E., and Fang, 8.0. 1962, Soil constituents and
properties in the adsorption of sulfate ions. Soil Sci. 94:276-
283.

. 1964. Iron and aluminum coatings in relation to sulfate
adsorption characteristics of soils. Soil Sci. Soc. Am. Proc.

28:632-635.

Chaudry, I.A. and Cornfield, A.H. 1966. The determination of total
sulfur in soil and plant material. Analyst 91:528-530.

Chesnin, L. and Yien, C.H. 1950. Turbidimetric determination of
available sulfate. Soil Sci. Soc. Am. Proc. 15:149-151.

Cortellessa, C.C. and Napoli, C.A. 1968. Potentiometric micro-
determination of the sulfate ion. Analyst 93:546-547.

Dean. C.A. 1966. A simple colorimetric finish for the Johnson-
Nishita microdistillation of sulfur. Analyst 91:530-532.

Dodgson, K. S. 1961 Determination of inorganic sulfate in studies
of enzymatic and non-enzymatic hydrolysis of carbohydrates and of
the sulfate esters. Biochem. J. 78:312-319.

Ensminger, L.E. and Freney, J.R. 1966. Diagnostic techniques for
determining sulfur deficiencies in crops and soils. Soil Sci.
101:283-290.

Fang, S.C., Nance, R.D., Chao, T.T., and Harward, M.E. 1962. A
chromatographic approach to the determination of sulfate
adsorption and exchange of less retentive soils. Soil Sci. 94:
14-18.

50

Forbes, E.A. 1973. The determination of microamounts of sulfate by
emission spectroscopy of barium with a nitrous oxide-acetylene
flame. Analyst 98:506-511.

Fox, R.L., Olson, R.A., and Rhoades, H.F. 1964. Evaluating the sulfur
status of soils by plant and soil tests. Soil Sci. Soc. Am. Proc.
28:243-246.

Freney, J.R. 1958. Determination of water-soluble sulfate in soils.
Soil Sci. 86:241-244.

Garber, R.W. and Wilson, C.E. 1972. Determination of atmospheric
sulfur dioxide by differential pulse polarography. Anal. Chem.

“431357-1359.

Goertzen, J.O. and Oster, J.D. 1972. Potentiometric titration of
sulfate in water and soil extracts using a lead electrode. Soil
Sci. Soc. Am. Proc. 36:691-692.

Gustafsson, L. 1960a. Determination of ultramicroamounts of sulfate
as methylene blue. II: The reduction. Talanta 4:236-243.

. 1960b. Determination of ultramicroamounts of sulfate as
methylene blue. I: The color reaction. Talanta 4:227-235,

Hesse, P.R. 1957. Effect of colloidal organic matter on the
precipitation of barium sulfate and a modified method for
determining soluble sulfate in soils. Analyst 82:710-712.

Hinze, W.L. and Humphrey, R.E. 1973. Spectrometric determination
of sulfate ion with barium iodate and the linear starch iodine
system. Anal. Chem. 45:814-815.

Hoeft, R.G., Walsh, L.M., and Keeney, D.R. 1973. Evaluation of various
extractants for available soil sulfur. Soil Sci. Soc. Am. Proc.
37:401-404.

Humphrey, R.E., Hinze, W.L., and Jenkines III, W.M. 1971.
Spectrophotometric determination of Hydrogen sulfide ion with
organic disulfides. Anal. Chem. 43:140-142.

Ignatova, V.P. and Abaturova, T.I. 1970. Use of the volumetric
titration method for determining sulfate ion is soils. Soviet
Soil Sci. pp. 117-120.

Iismaa, O. 1959. Microdetermination of sulfur in plant material.
J. Austr. Inst. Agric. Sci. 25:136-138.

Jenkinson, D.S. 1968. A titrimetric method for determining total
sulfur in mineral soils. Analyst 93:535-539.

Johnson, C.M. and Arkley, T.H. 1954. Microestimation of sulfur in
plant materials, soils, and irrigation waters. Anal.Chem.26:1525.

51

Johnson, C.M. and Nishita, H. 1952. Microestimation of sulfur. Anal.
Chem. 24:736-742.

Johnson, C.M. and Urlich, A. 1959. Analytical methods for use in
plant analysis. Calif. Agr. Exp. Sta. Bull. 766:57-58.

Khan, S.V. and Webster, G.R. 1968. An electrometric method for
determining water extractable sulfate in soils. Soil Sci. 105:

87-91.

Kirsten, W.I. and Patel, V.J. 1972. On the spectrophotometric
measurement of methylene blue. Michrochem. J. 17:277-284.

Lambert, J.L. and Zitomer, F. 1963. Simplified preparation of
cadmium iodide-linear starch reagent for colorimetric iodimetry.
Anal. Chem. 35:405.

Lysyj, I. and Zarembo, J. 1958. Rapid quantitative determination of
sulfur in organic compounds. Anal. Chem. 30:428-430.

Martin, J.M. and Stephen, W.I. 1967a. The nephelometric determination
of small amounts of sulfate ion. Anal. Chim. Acta 39:157-180.

1967b. An improved colloidal stabilizer for use in the
nephelometric determination of the sulfate ion. Anal. Chim. Acta

39:525-528.

Massoumi. A. and Cornfield, A.H. 1963. A rapid method for determining
sulfate in water extracts of soils. Analyst 88:321-322.

Reisenauer, H.M., Walsh, L.M., and Hoeft, R.G. 1973. Soils testing
and plant analysis, pp. 175. Soil Sci. Soc. Am., Inc., Madison,
Wisconsin, USA.

Roberts, S. and Koehler,F.E. 1968a. An X-ray spectrographic method
of determining total sulfur in soil extracts. Soil Sci. 106:164-
171.

1968b. Extractable and plant available sulfur in
representative soils of WashingtOn. Soil Sci. 106:53-59.

Ross, J.W. and Frant, M.S. 1969. Potentiometric titrations of sulfate
using an ion-selective lead electrode. Anal Chem. 41:967-969.

Saalbach, E., Kessen, G., and Judel, G.K. 1962. Investigations on the
determination of the plant-available sulfur content in soil.
Lanw. Frosch. 15:6-15.

Searle, P.L. 1968. Determination of total sulfur in soil by using
high-frequency induction furnace equipment. Analyst 93:540-545.

Shaw. W.M. 1959. Nitric-perchloric acid oxidation for sulfur in plant
and animal tissues. J. Agr. Food Chem. 7:843-847.

52

Souty, N. and Guennelon, R. 1967. I: Sulfur. Chemical (oxygen-bomb)
and X-ray fluorescence method of determination. Ann. Agrm. 18 (6)
653-667.

Spencer, K. and Freney, J.R. 1960. A comparison of several procedures
for estimating the sulfur status of soils. Austr. J. Agri. Res.
11:948-959.

Starkey, R.L. 1966. Oxidation and reduction of sulfur compounds in
soils. Soil Sci. 101:297-306.

Steinbergs, A. 1955. A method for the determination of total sulfur
soils. Analyst 80:457-461.

Steinbergs, A.. Iismaa, 0., Freney, J.R., and Barrow, N.J. 1962.
Determination of total sulfur in soil and plant material. Anal.
Chim. Acta 26:158-164.

Stephen, W.I. 1970. A new reagent for the detection and determination
of small amounts of the sulfate ion. Anal. Chim. Acta 50:413-422.

Tabatabai, M.A. and Bremner, J.M. 1970a. An alkaline oxidation method
for the determination of total sulfur in soils. Soil Sci. Soc.
Am. Proc. 34:62-65.

. 1970b. Comparison of some methods for the determination of
total sulfur in soils. Soil Sci. Soc. Am. Proc. 34:417-420.

Tiedemann, A.R. and Anderson, T.D. 1971. Rapid analysis of total
sulfur in soil and plant materials. Plant and soils 35:197-200.

Toennis, G. and Bakay, B. 1953. Photonephelometric microdetermination
of sulfate and organic sulfur. Anal. Chem. 25:160-165.

Varley, J.A. and Poon Yew Chin. 1970. Determination of water soluble
sulfate in acidic sulfate soils by atomic-absorption spectrOSCOpy.

Analyst 95:592-595.

Walker, D.R. 1972. Soil sulfateI. Extraction and measurement. Can.
J. Soil Sci. 52:253-260.

Walker, D.R. and Doornenbal, G. 1972, Soil sulfate II. As an index
of the sulfur available to legumes. Can. J. Soil Sci. 52:261-266.

Weisz, H. and Fritsche, U. 1970. Austauschreaktionen zur
nervielfachung von sulfid v. sulfat. Mikrochim. Acta 3:638-643.

Williams, C.H. 1967. Some factors affecting the mineralization of
organic sulfur in soils. Plant and soils 26:205-223.

Williams, H.R. and Burge, A.E. 1970. A method for the determination
of sulfur level in plant materials by atomic absorption
spectrophotometry. Qd. J. Agri. Anim. Sci. 27:169-175.

53

Williams, C.H. and Steinbergs, A. 1959. Soil sulfur fractions as
chemical indices of available sulfur in some Australian soils.
Austr. J. Agri. Res. 10:340-352.

HICHIGRN STQTE UNIV. LIBRQRIES

 

lllllll lllllllliilllllllllllll illllllllllllllillllll
9 06

312 31 924180