(10) Gerischer, H., ANAL.CHEM.31, 33
(1959).
(13) Hodgman, C. D., ed., “Handbook of
Chemistry and Physics,” 41st ed.,
Chemical Rubber Publishing Co., Cleve-
land, 1959.
(14) Jahnke, E., Emde, F., “Tables of
Functions,” 4th ed., Dover Publica-
tions, New York, 1945.
(15) Kambara, T., Tachi, I., J . Phys.
Chem. 61, 1405 (1957).
(16) Kolthoff, I.,, M., Lingane, J. J.,
“Polarography, 2nd ed., Vol. I, Inter-
science, New York, 1952.
(17) Koutecky, J., “Soviet Electrochem-
istry,” Vol. I, 177, Consultant’s
Bureau, New Yor!;1961.
Use of EIectricaIIy Excited Oxygen for the Low
Temperature Decomposition of Organic Substances
CHESTER E. GLEIT and WALTER D. HOLLAND
Tracerlab, Division of Laboratory for Elecfronics, Richmond, Calif.
b A stream of oxygen excited by a
radio frequency discharge can be used
to decompose organic substances prior
to trace element analysis. Biological
tissue, graphite, filter paper, and ion
exchange resin have been oxidized
by this method. Rates of 1 gram per
hour can be achieved with a 300-watt,
13.56-Mc. oscillator. Temperatures of
less than 100’ C. can be maintained.
Radioactive tracer studies demonstrate
that 17 representative elements can be
quantitatively recovered after com-
plete oxidation of the organic sub-
strate.
M ANY TECHNIQUES are mployed t o
decompose organic samples prior
to elemental analysis ( 2 ) . Wet ash-
ing, used extensively with mixtures of
hot mineral acids, is generally tedious
and potentially hazardous. The anions
which are introduced tend to interfere
with subsequent analyses. As im-
purities in reagents produce a significant
problem in wet ashing, dry ashing is
generally recommended (‘7) for trace ele-
ment analysis. However, during dry ash-
ing, which requires heating the specimen
to temperatures in excess of 400” C., ele-
ments from the container and the
atmosphere may be introduced into the
sample. Certain classes of biological
material tend to froth and char. Others,
high in nitrogen, may ignite upon heat-
ing. hlany substances are converted
into refractory, difficultly soluble com-
pounds. Serious losses are caused by
diffusion and volatilization. As vol-
1454 ANALYTICAL CHEMISTRY
(18) Koutecky, J., Cizek, J., Collection
Czech. Chem. Commun. 22, 914 (1957).
(19) Macero, D. J., Rulfs, C. L., J.Am.
Chem. SOC.81, 2942 (1959).
(20) Matauda, H., 2. Elektrochem. 61,
489 (1957); 62, 977 (1958).
(21) Matauda, H., Delahay, P., J. Am.
Chem. SOC.82, 1547 (1960).
(22) Matauda, H., Oka, S., Delahay, P.,
Ibid 81, 5078 (1959).
(23) &holeon, M. M., Ibid., 76, 2539
(1954).
(24) Oldham, K. B., J . Electrochem. SOC.
107, 766 (1960).
(25) Randles, .J; E. B., “Progress in
Polarography, P. Zuman, ed., Vol.
I, p. 123, Interscience, New York,
1962.
(26) Randles, J. E. B., Trans. Faraday
SOC.44, 327 (1948).
(27) Reinmuth, W. H., AKAL. CHEM.33,
485, 1438 (1961).
atility depends on the oxidation state
of the element and the composition of
the substrate, one cannot accurately
predict which elements will be lost in a
specific sample.
Recently, Schoniger’s (6) modification
of the Hempel oxygen flask technique
(8) has been widely used for micro-
analysis. Because of the high pressure
developed in the flask, this technique is
limited t o small samples.
An alternate method, described in this
paper, employs a high frequency elec-
tromagnetic field t o produce a stream
of reactive oxygen, which decomposes
the organic substances. Superficially,
the apparatus employed in this method
is similar to high temperature dry ash-
ing in an induction heated tube furnace.
Both methods employ a n essentially
closed system to minimize the introduc-
tion of atmospheric impurities and per-
mit recovery of volatilized elements.
However, induction heating units in-
corporate a metallic substance into the
sample or combustion boat to heat the
sample to high temperature and provide
the requisite activation energy. On the
other hand, in the radio frequency dis-
charge method, electrical energy is trans-
ferred directly to a stream of low pressure
gas, producing highly excited states of
oxygen. The reaction of these species
with the elements in the organic sample
is selective, minimizing general heating.
Because the sample temperature is low,
volatility and diffusion losses are sub-
stantially decreased. Nonvolatile
mineral constituents remain in the
(28) Ibid., p. 1793.
(29) Reinmuth, W. H., J . A m . C h e m SOC.
79, 6358 (1957).
(30) Reinmuth, W. H., J. Phys. Chem.
61, 1405 (1961).
(31) Sevcik. A,. Collection Czech. Chem.
‘ commun.’l3, ’349 (1948).
(32) Vernotte, P., “ThBorie et Practique
des Shies Divergentes,” Publications
Scientifique du Ministere l’air) KO.207,
Paris, 1947.
(33) Vielstich, W., Delahay, P., J . S m .
Chem. SOC.79. 1874 (1957).
(34) Wijnen, ~. D., ’Rec. T r m . Chim.
79, 1203 (1960).
(35) Wilson, E. B., “An Introduction t o
Scientific Research,” McGraw-Hill, Xew
York, 1952.
RECEIVEDfor review June 11, 1962.
Accepted August 8, 1962.
sample contiiiner in a form amenable to
subsequent quantitative determination.
EXPERIMENTAL
Apparatus. Figure 1 illustrates
schematically the principal apparatus
employed in this study. Molecular
oxygen (Linde commercial grade)
passes through a flow meter and is
admitted t o the system through a
needle valve, A . T h e oxygen flows
through a borosilicate glass reaction
chamber, B , 30 cm. long and 4 cm. i n
diameter. T o determine the effects of
gas flow rate, pressure, and power, speci-
mens were placed in a borosilicate glass
boat in chamber B, 25 cm. from the
oxygen inlet. Volatility and recovery
experiments employed a 30-cm. long
water-cooled chamber, C. At a point 15
cm. from chamber B the inside diameter
of the tube was reduced from 4 cm. to 1.2
cm. To perform volatility studies,
samples were placed in a n open glass
boat 20 cm. from chamber B. To
determine the types of materials nThich
could be ashed by this technique, speci-
mens were placed in chamber C in
glass tubes 2.5 cm. in diameter.
Exhaust vapors pass through a cold
trap, D,surrounded by a mixture of dry
ice and acetone. A general purpose
mechanical vacuum pump (Kinney
model KC-3) is attached a t E. To
protect the Tygon tubing connections
from the action of active oxygen, a
small piece of platinum gauze is placed
in tube E. A mercury manometer is
attached to sidearm F .
Power is supplied by a conventional
radio frequency oscillator, G. The
oscillator has a crystal-controlled driver
 B F

L-

Figure 1, Low temperature, radio frequency oxidation

m
a p pa ratus
o \ g

which operates a power stage employing

an Eimac 4-125A power tetrode (Eitel-
RlcCullough, Inc.). T o comply with
Federal Communications Commission
radiation requirements on Industrial,
2ol 0.5 I I5 2 25
Scientific, and Medical equipment, with HOURS
minimal shielding, a frequency of 13.56 Figure 2. Effect of surface area on oxidation
M e . was chosen. Inductive coupling rate
transfers power to the reaction chamber.
Coil H consists of 18 turns of 1/4-inch Abscissa indicates weight of graphite oxidized during prior
copper tubing. The plate tank is tuned 30-minute interval
by means of a variable capacitor, 1.
K i t h a coil having a n inductance of
10 phenries, a 50- to 150-picofarad air was ashed in a similar manner. The rate is more complex. For each reac-
capacitor is satisfactory. experiments with iodine employed tion tube, a distinct optimum pressure
W h ~ t r I mNo. 42 filter paper disks 8s was found, The maximum occurs a t a
the substrate. pressure of 1.5 mm. of mercury for the
PROCEDURE
30 cm. X 4 cm. chamber.
Measurements of the effect of pres- RESULTS The relationshir, between oxidation
sure, flow, and power on oxidation rate rate and sample position is strongly
employed 0.25-inch lengths of 0.242 A wide variety of organic substances
ashed, including muscle tissue, fat, dependent on flow and pressure. The
inch in diameter spectroscopic grade
graphite electrodes weighing 300 mg. fecal matter, ion exchange resin (Dowex highest rates were recorded for samples
(Sational Carbon Co., KO. 3829). AG50W, 200- to 400-mesh), cellulose mounted axially within several centi-
Graphite felt (National Carbon Co., and poly(viny1 chloride) filter paper, meters of the downstream end of the
Grade WDD) was used to obtain a activated charcoal, and a 40-gram rat. coil. At the upstream end of the coil,
larger surface to weight ratio. Speci- I n all cases, the ash was completely rates were approximately 30y0lcwer. At
mens were weighed periodically to a power of 250 watts, pressure of 1.0 mm.
determine oxidation rate. A radio- soluble in mineral acid, and the weight
of ash was the value anticipated from of mercury, and flow rate of 10 cc. per
active tracer, NaZ2,was added to several minute, graphite rods could be oxidized
specimens to determine that the change muffle furnace ashing and known
in weight was not caused by graphitic mineral content. at a rate of 35 mg. per hour, in a trap
particles being blown from the combus- Figure 2 shows the reaction rates for surrounded by dry ice 40 cm. down-
tion boat. Power was calculated from graphite rod, carbon felt, and a graphite stream of the coil. I n the absence
the measured cathode current and plate rod that had been shaved into small of the dry ice coolant, a rate of only 3
voltage. Plate loss was calculated from slivers. Oxidation rate is a function of
known circuit parameters and agrees exposed surface and decreases markedly
with visual observation of plate tem-
perature. as the surface becomes covered with a Table 1. Effect of Flow, Pressure, and
The majority of volatility experi- mineral residue. The change in de- Power on Oxidation Rate of Spectro-
ments employed 1.0-ml. aliquots of composition rate of tissue samples with scopic Grade Graphite Rods
whole human blood containing 20% time is similar to that shown for carbon
acid citrate as an anticoagulant. Radio- felt. Oxide
active tracer, 100 pl., was added to the tion
Table I presents the effect of pressure, Flow Pressure, rate,
blood. After 24 hours, the samples flow, and power on oxidation rate. rate, mm. of, Power, mg:/30
mere dried in a desiccator containing Over the range of 100 to 400 cc. per cc./min. Hg wattts min.
CaC12, placed in chamber C and exposed
to a stream of activated oxygen for 1.5 minute (S.T.P.), the rate of decomposi- 10 5 .O 250 25
hours. -4pressure of 400 microns of mer. tion is essentially independent of the 18 5.0 250 45
oxygen flow rate (Figure 3). Rate 25 5.0 250 86
cury, a flow rate of 4 cc. per minute, and a 50 5.0 250 94
delivered pon'er of 150 watts were em- increases linearly with power in the 100 5.0 250 114
ployed. RIaximum sample temperature range of 120 to 300 watts. Doubling 200 5.0 250 116
n a s less than 100' C. iifter ashing, the power produces a 25y0 increase in 300 5.0 250 120
the combustion boat was reweighed and oxidation rate. For specimens with 400 5.0 250 109
counted in a Tracerlab SC-57 well- 70 5.0 120 75
larger surface area, the change in rate 70 5.0 165 95
counter. The contents of the boat were with increased power is greater. The
dissolved in a mineral acid and re- 70 5.0 230 100
discharge, which is characterized by an 70 5.0 300 122
counted in a conventional 4-ml. vial. 70 0.8 250 165
All glassware was rinsed with an intense light blue glow, could not be
70 1.3 250 213
appropriate mineral acid, and the maintained when less than 120 watts 70 3.3 250 135
wash solutions were counted. A sample was delivered to the induction coil. 70 5.3 250 123
of alfalfa grown in soil enriched in Se76 The relationship between pressure and

VOL 34, NO. 1 1 , OCTOBER 1962 1455

 mg. per hour could be attained a t this I I
I I
I I
I 1
I I
position.
Table I1 presents the results of the
volatility studies. For comparison, the
data obtained in a recent study (4)
of the retention of trace elements in
muffle furnace ashing are given. Of
the 17 elements tested, only four were
partially volatilized. The mercury,
silver, and gold activities were located by
a r-ray sensitive survey meter on the
vessel walls a short distance on both sides
of the combustion boat. Between 1 and 0
9% of the iodide activity was deposited
on the walls of the reaction vessel. The 20
remainder was found in the cold trap,
which had been filled with activated 0 1 I I I I I I I
charcoal for the iodine experiments. 0 100 200 300 400
Large variations in the distribution of FLOW R A T E , C c l n i n .
radioactive tracer between the sample
container and other parts of the system Figure 3. Effect of oxygen flow rate on the oxidation rate of graphite
occurred in replicate experiments with rods at a constant pressure of 5 mm. of mercury and a delivered power
the four volatile species, Agf, Auf3, of 250 watts
Hgf2, and I-. Average values are
given in Table I1 for these species. fractionation due t o a portion of the are expected on the basis of well known
Volatilization loss occurred concur- sample remaining in a partially oxidized, gas discharge phenomena. At low
rently with sample oxidation. In a carbonaceous residue should be obvi- pressure most of the energy imparted to
typicalexperiment with a silver chloride- ated. the gas is carried off as kinetic energy by
blood mixture, oxidation was 96% The data on oxidation rate have a free electrons. At higher pressures a
complete after 10 minutes. During this precision of 4%, as determined from greater portion of energy is transferred
period, 26% of the silver activity had replicate experiments. The slow rise to the atomic species, and a t high pres-
been volatilized. I n the following 120 in rate with power suggests that the sures recombination and de-excitation
minutes less than 1% of the Agllo re- arrangement employed does not fully reactions between species are more
maining in the combustion boat was utilize the excited oxygen species. frequent. When specimens are placed
transported to other parts of the sys- That the intensity of the blue glow is at a great distance from the induction
tem. not appreciably decreased when the gas coil, a low pressure is desirable to reduce
passes over small organic samples recombination reactions. The unex-
DISCUSSION supports this conclusion. Higher rates pected increase in rate obtained by
can be achieved without increase in placing the sample directly in a cold
The radio frequency discharge tech- power by a more complex electrical trap 40 cm. from the coil results from the
nique offers a convenient method of generator and reaction chamber. Be- suppression of de-eucitation reactions
decomposing a wide range of organic cause the organic sample does not ap- on low temperature surfaces. The in-
substances. No variation of oxidation preciably load the plate circuit, re- tensity of the blue glow is a good ap-
rate with type of organic compound was turning the capacitor during ashing is proximation of oxidizing power. In the
noted, indicating that the metastable not required and unattended operation cold trap experiment, the gloK n a s
oxygen species formed by the discharge is feasible. visible a t the base of the trap only
has sufficient energy to rupture all The maxima observed in the studies when the coolant mixture was present.
C-C and C-H bonds. Therefore, of the effect of flow rate and pressure Blood was chosen as the substrate
in the volatility evperiments to provide
a direct comparison with the standard
muffle furnace technique. Whole blood
Table II. Recovery of Radioactive Tracers in Dry Ashing is rich in chelating agents and chloride
(Tabulated values are per cent recovered from vessel) ions known to increase markedly the
R.f. discharge, 1.5 hr. Muffle furnace volatility of many cations. Although
Trap and 24 hr., 3 hr., volatility loss is appreciably less in the
Sample Boat chambers 400” C. 900’ C. radio frequency discharge method, the
++
SbC4 blood
HAEO~ blood
99
100
0
0
67
23
9
0
difference in retention of iodide and
iodate indicates that reproducible losses
++
CsCl blood
COC12 blood
100
102
0
0
...
98 30
mag occur in both methods. A further
reduction in volatility can be obtained
+
CuClz blood
+
CrCla blood
101 0 100
99
58 by oxidizing the sample directly in the
100 0 56
+++
AuCL blood 70 30 19 0 cold trap. The use of liquid nitrogen
NaI filter paper 31 69 ... .. as a coolant is not advisable because of
NaI03 filter paper 100 0 ... .. the accumulation of liquid ozone and
+
FeC13 blood 101 0 86 27
+ ++
Pb(N0a)z blood 100 0 103 13 other unstable species a t low tempera-
MnClz blood 99 0 99 79 tures.
Hg(N0a)z blood 92 8 <1 0 The low volatility of elements such as
(“4)&fOoi
Alfalfa
+ blood 100 0 100
...
83
..
arsenic and lead is surprising. Most
99 0
+
AgCl blood 72 28 65 21 inorganic compounds of these species are
NaCl ++ blood 100 0 ... .. reported (5) to be more volatile than
those of silver in an arc discharge.
ZnC12 blood 99 0 100 30
The muffle furnace study (4)also indi-

1456 ANALYTICAL CHEMISTRY

cates that arsenic(II1) is more volatile
than silver(1) a t all temperatures.
“Loosely bound” arsenic in blood is
reported (1) to volatilize as a n unknown
compound a t 56” C.
ACKNOWLEDGMENT
We acknowledge the valuable advice
of Robert M. Main. The technical
assistance of William Mills, Robert
Wrigley, and Mrs. Ollwyn Brothers is
Polarographic Investigations in the Ammoniate of
Sodium Iodide
D. E. SELLERS’ and G. W. LEONARD, Jr.2
Department of Chemistry, Kansas State University, Manhattan, Kan.
b The preparation and use of the
ammoniate of sodium iodide for a
polarographic solvent has been de-
scribed. Measurements on both in-
organic and organic compounds have
been accomplished. In most instances,
the results indicate normal polaro-
graphic behavior.
T HE INVESTIGATION of the use of
various inorganic ammoniates as
suitable polarographic solvents has be-
come more and more frequent ( I , 5, 7‘).
These systems have certain advantages
over other nonaqueous media in that
there is no interference from oxygen,
these systems are highly ionized result-
ing in low internal cell resistance, and
large numbers of both inorganic and
organic compounds which are polaro-
graphically reducible may be investi-
gated.
The purpose of this investigation was
to study the KaI.(iTH3)1 solvent with
respect to reproducibility in solution
preparation, finding a suitable reference
electrode, to characterize the usable
polarographic range, and to note the
effect of concentration on both the dif-
fusion current and half-wave potentials
and other pertinent factors which were
considered necessary for the character-
ization of this solvent for polarographic
use.
EXPERIMENTAL
Apparatus. The electrolysis vessel
consisted of the upper 3 inches of a
large 25 X 100-mm. test tube fitted
1 Present address, Department of Chem-
istry, Southern Illinois University, Car-
bondale, Ill.
a Present address, Code 452, Naval
Ordnance Test Station, China Lake,
Calif.
greatly appreciated. The alfalfa grown
in Se75was obtained through the cour-
tesy of C. Johnson, University of Cali-
fornia, Davis, Calif.
LITERATURE CITED
(1) AhAens, L. H., “Spectrochemical Anal-
yam, p. 72, Addison-Wesley, Cam-
bridge, Mass., 1950.
(2) Dunlop, E. C., in Kolthoff, I. M.,
Elving, P. J., Sandell,.E. B., “Treatise
on Analytical Chemistry,” Part I,
Chap. 25, Interscience, New York, 1961.
with a three-holed rubber stopper
which held the dropping mercury
electrode, a glass-encased silver ref-
erence electrode, and a n ammonia
exit tube which was connected to a
metallic sodium drying tube. A Sar-
gent Polarograph Model X I I , coupled
with a wave spreader (2), was used to
record the current-voltage curves. Cali-
bration points were marked on the po-
larographic record by adjusting the
galvanometer a t 20 and opening the
shutter a t the beginning and end of
each polarogram. A span of 0.2 volt
was used for the determination of the
half-wave potentials except when it was
impossible to record the entire electrode
process when a span of 0.5 volt was
used. A Rubicon portable potentiom-
eter was used to measure the voltage
a t the beginning and end of each
polarogram a t the calibration points.
The resistance of the cell circuit,
measured with an Industrial Conduc-
tivity Bridge Model RC, was 100 ohms.
The temperature of the sample being
polarographed was maintained a t 24” =t
0.2” C. with a Sargent constant tem-
perature bath.
The dropping mercury electrode pos-
sessed a drop time of 4.98 seconds a t
-0.484 volt us. the silver-silver iodide
electrode and its capillary constant
was 1.47 mg.2/3t-1/2. The silver-silver
iodide reference electrode was made by
taking a 6-inch piece of pure silver
wire, 1 mm. in diameter, and making a
six-turn coil with an outside diameter
of approximately 3 mm. This coil was
then encased in small-bore glass tubing
in which a small hole was placed just
below the rubber stopper. After being
cleaned with dilute nitric acid, it was
electrolyzed a t -3.0 volts for approxi-
mately 1.5 hours, rinsed, and stored in
distilled mTater for future use. The
area of the electrode which was in con-
tact with the solution was approximately
1.5 sq. cm. A second reference elec-
trode prepared in the same manner
agreed with the first within 5 mv.
VOL. 34, NO. 1 1, OCTOBER 1962
(3) Hempel, W., Z. Angew. Chem. 13, 393
(1892).
(4) Pijck, J., Gillis, J., Hoste, J.. Intern.
J . Appl. Radiation Isotopes 10, 149
(1961).
(5) Satterlee, H. S., Blodgett, G., IND.
ENG.CHEM.,ANAL.ED. 16, 400 (1944).
(6) Schoniger, W., Mikrochim. Acta, 123
(1955); 869 (1956).
( 7 ) Thiens, R. E., in Glick,. D., “Methods
of Biochemical Analysis," chap. 6,
Interscience, New York, 1952.
RECEIVED for review July 5, 1962. Ac-
cepted August 6, 1962.
Reagents. Sodium iodide, C.P. (J.
T. Baker Chemical Co.) was satis-
factory for use without further purifi-
cation: It yielded a clear solution
upon condensation with anhydrous
ammonia (-4rmour and Co.) and gave
no reducible species within the usable
polarographic range. Sodium iodide
from other sources left a small white
residue after complete condensation.
To remove the last trace of moisture,
the ammonia was passed through a 6- to
8-inch tube containing metallic sodium.
The substances being analyzed polaro-
graphically were either reagent grade
or prepared and purified by methods
described in the literature.
Procedure. Using a microbalance,
weigh into the polarographic cell
the required amount of solute. If
necessary, add approximately 20 mg.
of gelatin as a maximum suppressor.
Add approximately, accurately weighed,
10 grams of sodium iodide (previously
dried a t 110” C. for a t least 1 hour)
to the cell and stopper the cell with a
rubber stopper until the polarographic
solution is prepared. Place the cell
in the water bath maintained a t the
proper temperature, and insert the
rubber stopper containing the bubbler
and exit tubes. Pass dry ammonia gas
through the cell until complete con-
densation has occurred, approximately
15 to 30 minutes. Remove the bubbler
system and immediately insert the
D.il1.E.-reference electrode unit. Re-
cord the current-voltage curve over the
desired range, making sure that the
reference points and potentials are
recorded.
RESULTS A N D DISCUSSION
Stability and Composition Studies
of Various Ammonia Systems. T o
describe these solutions more com-
pletely, i t was advantageous t o charac-
terize them with respect t o their
stabilities and compositions. Because
0 1457