THE DETERMINATION OF OXYGEN CONCENTRATION IN SOIL WITH COLLODION COATED SILVER WIRE ELECTRODES Thesis for the Degree of Ph. D. , MICHIGAN STATE UNIVERSITY TSE CHUN YANG 1970 .-. 4" .y a «Sigmund-urn» "‘J LIBRARY I" Miclfigan State University 1HESIS This is to certify that the thesis entitled THE DETERMINATION OF OXYGEN CONCENTRATION IN SOIL WITH COLLODION COATED SILVER WIRE ELECTRODES presented by Tse Chun Yang has been accepted towards fulfillment of the requirements for PhoDo degree in SOil SCienCe Date June 10, 1 0—169 y umsmc IY IIOAB & SONS' BOOK “In!" INC. l IIIADV ”Inf-s a'. . s...“ "fi‘ifi-IIII 'I .’ LIBRARY Michigan State University THESIS This is to certify that the thesis entitled THE DETERMINATION OF OXYGEN CONCENTRATION IN SOIL WITH COLLODION COATED SILVER WIRE ELECTRODES presented by Tse Chun Yang has been accepted towards fulfillment of the requirements for PhoD. degree in 8011 SCienCe Date June 10, 1 0-169 if IIN‘DING IV TIME 3. sons' your mum me 0-.A-V ‘uflqu. ABSTRACT THE DETERMINATION OF OXYGEN CONCENTRATION IN SOIL WITH COLLODION COATED SILVER WIRE ELECTRODES BY Tse Chun Yang An instrument was designed to measure oxygen concentra- tion in soil. It is an amperometric measurement which uses a silver microelectrode coated with collodion as cathode and a silver-silver chloride anode. An integrated current of short duration is measured shortly after closing the circuit. The current is shown theoretically and experimentally to be related with oxygen concentration of the medium. Oxygen concentration is expressed in terms of oxygen saturation thus eliminating temperature correction. Consistent readings were obtained when measurements were taken from a silt sample in 0.02 N KCl at 70 cm tension. The meter is shown to be useful to measure oxygen concentration in soil at or near saturation. THE DETERMINATION OF OXYGEN CONCENTRATION IN SOIL WITH COLLODION COATED SILVER WIRE ELECTRODES BY Tse Chun Yang A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of CrOp and Soil Sciences 1970 To My Parents and Wife ii ACKNOWLEDGMENTS The author wishes to express his sincere gratitude to Dr. A. E. Erickson for his enthusiastic guidance throughout this work. Hearty thanks are extended to Dr. R. L. Cook and the late Dr. C. Y. Sheng; with their encouragement the study was realized. Sincere appreciations are also due to his guidance members: Dr. A. Timnick, Dr. C. Cress, Dr. M. M. Mortland, Dr. B. G. Ellis and Dr. R. J. Kunze. The financial assistance provided by the Taiwan Agri- cultural Research Center and National Science Council of Republic of China is gratefully acknowledged. iii TABLE OF CONTENTS Page I . IMRODUCTION O O O O O O O O 0 O O O O O O O O O 1 II. REVIEW OF LITERATURE. . . . . . . . . . . . . . 2 Instruments for Oxygen Measurement. . . . . 2 Oxygen Reduction at a Metal Cathode and Factors Affecting It. . . . . . . . . . 8 III. THEORETICAL ANALYSIS. . . . . . . . . . . . . . 12 IV 0 TIE POROUS COATING I I O O O O I O O O O I O O O 18 Determination of the Thickness of Coating . 19 Estimation of the Diffusion Coefficient in the coating 0 O O O O O O O O O O O O O 21 A Standardization Process . . . . . . . . . 22 V. THE OXYGEN CONCENTRATION METER. . . . . . . . . 24 The Electrode System. . . . . . . . . . . . 24 The Timing Relays . . . . . . . . . . . . . 29 The Integrator. . . . . . . . . . . . . . . 29 The VOltmeter . . . . . . . . . . . . . . . 50 VI. RESULTS AND DISCUSSION. . . . . . . . . . . . . 32 The Collodion Coating . . . . . . . . . . . 35 The concentration ratio . . . . . . . . 34 The diffusion coefficient in the coat- ing . . . . . . . . . . . . . . . . 54 The choice of the steady-state time . . 55 Testing the thickness of coating. . . . 55 The costing thickness and sensitivity . 41 The Effect of Applied E.M.F.. . . . . . . . 45 The Effect of Salt Concentration. . . . . . 44 Oxygen Concentration Measurements in Flooded and Wet Soils . . . . . . . . . 48 iv TABLE OF CONTENTS--continued Page VII. SUMMARY AND CONCLUSIONS. . . . . . . . . . . . . 54 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . 56 APPENDICES A. Description of the Operating Controls and Terminals of the Oxygen Concentration Meter . . 60 B. Operating Instructions of the Oxygen Concentra- tion Meter . . . . . . . . . . . . . . . . . . . 64 TABLE LIST OF TABLES Page Comparison of the standardized currents from air saturated media as measured by electrodes of various coating thicknesses. . . . . . . . . 58 Analysis of variance for the standardized cur- rents measured in a silt sample leached with air saturated 0.02 N KCl and equilibrated at 70 cm tension using electrodes of various coat- ing thicknesses . . . . . . . . . . . . . . . . 59 Differences in the standardized currents from electrodes of various coating thicknesses as measured in a silt sample leached with air- saturated 0.02 N KCl and equilibrated at 70 cm tension . . . . . . . . . . . . . . . . . . . . 40 Oxygen concentrations and oxygen diffusion rates in the top soil (2-4 cm) at the Univer- sity's Soils Farm . . . . . . . . . . . . . . . 52 vi LIST OF FIGURES FIGURE 1. Block diagram of the oxygen concentration meter. 2. Oscilloscope photo showing current-time rela- tionships after a -0.65 v emf was applied to a collodion coated silver electrode. . . . . . . 5. Standard curves obtained from electrodes of various coating thicknesses. . . . . . . . . . . 4. Oxygen diffusion current at various applied emf. 5. Oxygen diffusion currents measured in KCl solu- tions at air saturation. . . . . . . . . . . . . 6. Oxygen diffusion currents measured at various added resistances. . . . . . . . . . . . . . . . 7. Oxygen concentration in sieved soils after being submerged under air-saturated 0.02 N KCl . . . 8. Pictorial diagram of the oxygen concentration meter 0 I O O O O O O O O O O O O O O I O O O 0 O 9. Wiring diagram of the oxygen concentration meter vii Page 25 36 42 43 45 46 50 61 62 I. INTRODUCTION It is well-known that an oxygen deficient soil sup- presses the growth of most plants and their associated soil microorganisms. Monitoring soil oxygen status in the rhiZOSphere of poorly aerated soil is an important soil measurement for studying crop production. Two parameters are needed for the complete exPression of the soil oxygen status: oxygen concentration and oxygen flux. The oxygen status of a soil, due to the microhetero- geniety of the soil, is not uniform. In a location it may differ greatly between two peds from that within the peds. Therefore, soil oxygen status described in micro-terms would be more meaningful than in macro-terms. Soil oxygen flux has been successfully measured,l since Lemon and Erickson (1952) introduced the oxygen diffusion meter. The use of platinum microelectrode provided good spatial resolution of oxygen flux. The oxygen flux as seen by a fine root is closely simu- lated by a thin wire electrode. However, an apparatus for determining dissolved soil oxygen concentration of comparable resolution has not yet been realized. The purpose of this study is to investigate the possibility of constructing such an instrument. J'Under favorable conditions described on page 5. II. REVIEW OF LITERATURE Numerous writers have discussed the importance of soil oxygen to plant growth and to soil chemical reactions. Some reviews of the literature are Bergman (1959), Black (1968), Brandt (1965), Grable (1966), and Russell (1952). In this study, therefore, the focus is on the develOpment of instru- ments for measuring oxygen concentration and factors affect- ing such measurements. Some of the instruments are not for soil use, but they certainly play a role in the sequence of development. The principle and design of these instruments may enlighten ideas for further improvement. INSTRUMENTS EOR 0mm MEASUREMENT In 1897, Danneel (11) found that current passing through two platinum electrodes in aqueous solution depends on oxygen concentration. Precise measurement of dissolved oxygen was later achieved by using a dropping mercury electrode or a rotating electrode (Kolthoff and Lingane, 1952). The applica- bility of stationary electrodes to i§_situ and ip_vivo measure— ment received, naturally, more attention from biologists and related scientists. Blinks and Skow (1958) used a stationary platinum elec- trode to measure oxygen concentration change at the surface of green leaves in solution. Lemon and Erickson (1952) introduced the platinum microelectrode method to measure soil oxygen diffusion rates and this method has been widely adopted. Satisfactory results can be obtained when a clean electrode is completely covered with soil solution and the cell resistance is within a favorable range which is in the range of soil oxygen deficiency to plants (Van Doren and Erickson, 1966). Davies and Brink (1942) constructed a recessed micro- electrode for measuring oxygen concentrations in the cat brain. The platinum wire was sealed 1 mm inside the opening of a glass tube. This construction for oxygen concentration measurement was supported by a sound diffusion theory treatment. In blood oxygen measurements, the undesirable effect of blood cells was minimized by coating a platinum electrode with collodion (Drenckhahn, 1951) or cellophane (Clark, Wolf, Granger, and Taylor, 1955). Clark (1956) enclosed a platinum cathode and a silver anode together behind a polyethylene membrane for blood oxygen determination. Material such as teflon (Sawyer, George, and Rhodes, 1959; Krog and Johansen, 1959) and silicon rubber (Flynn, Kilburn, Lilly, and Webb, 1967) are also used as mem- branes. These membranes are permeable to oxygen but restrict other substances from diffusing through them. Poisoning can thus be minimized and electrode stability greatly improved. Another outstanding feature of the membrane covered electrode is that regardless of whether measurements are made in gas or in liquid phase the readings are almost the same, provided both are at the same oxygen tension (Reeves, Rennie, and Pappenheimer, 1957; Krog and Johansen, 1959). Carritt and Kanwisher (1959) incorporated a thermister in the electrode system to compensate for the high tempera- ture dependence of the diffusion rate (8.5% per degree at 25°C for polyethylene) in the cell. McIntyre and Philip (1964) used this electrode system in their apparatus to study gas diffusion into soil. Mancy, Okun and Reilley (1962) felt that a single fixed thermister p§£_§g cannot offer adequate temperature compensation because the permeability coefficient of plastic sheets vary from one batch to another (membrane requires frequent replacement) and because of variable distor- tion applied to the membrane on mounting. They provided nomograph charts for temperature correction with their oxygen analyzer which has a silver cathode and a lead anode in 1 M KOH. The spontaneous cell reaction eliminates the requirement of an external current source. Current from a membrane- covered electrode is independent of the variation in oxygen solubility due to types and concentration of salt in the medium measured. All the membrane covered cells require a minimum amount of stirring while taking measurements, except in the case of gases which mix readily. Insufficient stirring gives readings dependent on the diffusion impedence of the medium measured. Enoch and Falkenflug (1968) suggested to use an additional sheet of membrane to enclose a thin gaseous volume on the end of a Beckman oxygen sensor (no. 59065). This volume which is in equilibrium with the soil environment through a greater surface than with the original effective surface of the electrode, in effect reduces the diffusion impedence effect of the environment. The time required to reach a steady current reading increases with the spacing between the two membranes. Some stirring was still applied by bubbling the solution with gas in their experiment. A good membrane- covered sensor to be used in soils should, at least, give consistent readings when it is placed in gas and in unstirred aqueous solution of the same oxygen tension. Willey and Tanner (1965) designed a rugged membrane- covered electrode with temperature compensation for measuring oxygen concentration in soil. Steady readings can be attained 2 1/2 minutes after the circuit is closed. The electrode, when placed in an access tube, has a diameter of about one inch. Because of its large size it may be difficult to use in the field to measure the microenvironment of roots. However, it still seems to be the best and most compact tool available for soil oxygen concentration investigations. Brandt (1955) attempted to use a bare platinum wire electrode to estimate oxygen concentration in bentonite sus- pensions and in soil. Oxygen concentration was calculated from the transient currents, recorded photographically from an oscillograph trace, after a few tenths of a second of circuit closure. Using the diffusion current after a short duration of circuit closure virtually limits the diffusion path. Only the diffusion character in the immediate vicinity of the electrode surface is important. Good agreement was attained between the measured and actual oxygen concentration of a bentonite suspension; but the results were erratic when soil oxygen concentration was measured. The variability due to the platinum electrode itself, as pointed out by Brandt, was later successfully overcome by using silver electrodes (Erickson, Fulton and Brandt, 1964; Wu, 1967). Wu (1967) used a set of timers and an integrator to substitute for the time-consuming oscilloscope photography for reading transient current. Reasonably consistent readings were obtained from saturated sands and sandy loams of the same oxygen concentra- tion. But variations were found at higher moisture tensions, around 10 cm for sands and 50 cm for sandy loams, and high electrical resistances, approximately 50 kilo-ohms between the Ag and Ag-AgCl. Both Brandt and Wu assumed that the diffusion coefficient of oxygen in soil in the immediate vicinity of the electrode is constant. The validity of this assumption was not explained. Non-polarographic methods for determining oxygen concen- tration in unsaturated soil are also available. Hutchins (1921) determined oxygen supplying power by burying a porcelain cup in soil and flushing it with nitrogen gas. Oxygen in the outflowing gas was absorbed by a solution of pyrogallol and potassium hydroxide, and determined colorimetrically. The apparatus is too clumsy for field use. A relatively compact diffusion tube was designed by Raney (1950). It has a lateral port and a valve at the lower end. After 10 minutes of diffusion the gas in the chamber was analyzed by a Beckman model D oxygen analyzer. Similar kinds of appara- tus were also described by Taylor and Abraham (1955) and van Bavel (1954). This could be used for oxygen concentra- tion determination if the time of diffusion was long enough to allow equilibrium to be established. Hack (1956) was able to take a small soil air sample, as little as 0.01 cc, by inserting a microsyringe with a lateral port into the soil. Small samples and large samples were compared. Results showed that small samples are more sensitive to oxygen concentration changes in soil than are large samples. Yamaguchi, Howard, Hughes, and Flocker (1962) used a syringe to take air samples from a diffusion chamber so that sampling depth is not limited by the length of the needle. No apparatus is available to successfully detect oxygen concentrations in saturated and unsaturated soils with a micro-spatial resolution similar to the plant root environ- ment. OXYGEN REDUCTION AT A METAL7CATHODE AND FACTORS AFFECTING IT Laitinen and Kolthoff (1941) showed the presence of H202 as a product of electrolysis of oxygen at a platinum electrode in KN03 solution. Lingane (1961) reported that direct reduc- tion of oxygen at a platinum cathode is diffusion controlled in the presence of a small amount of platinum oxide (about one-tenth the amount on a fully oxidized electrode is suffi- cient), but kinetically controlled if the oxide is not suffi- cient. He also concluded, by using an anodized platinum electrode in 1 M H2804 and 1 M NaOH, that in both acid and basic aqueous media, oxygen was reduced all the way to water rather than to H202. At the same time, Sawyer and Interrante (1961), working on both oxidized and reduced electrodes of various metals including Pt, Pd, Ni, Ag, Au, Ta, W, Cu, and Pb, pointed out that the reduction does not involve H202 at the oxidized electrodes but does produce H202 at the reduced platinum electrode. Reaction mechanisms were also proposed, i.e., at an oxidized electrode, Pt(0H)2 + 2e- ——>Pt + 20H— 2Pt + 02 + 2H20 —> 2Pt(OH)2, and at a reduced electrode, Pt(0H2)n + 02 + e’ —>-Pt(OH2)n_1(0H)- + H02 slow Pt(OH2)n_1(0H)_ + H2O —> Pt(0H2)n + 0H- fast 2H02 : H202 + 02 fast. Later Sawyer and Day (1965) concluded that the mechanism at the oxidized electrode is dependent upon the electrode material, whereas the mechanism at the reduced electrode is largely independent of electrode material. The cyclic reac- tions of platinum hydroxides coincides with Lingane's conclu- sion, i.e., if the platinum electrode does not have enough oxide, the reaction becomes rate limiting rather than diffu- sion controlled. The presence of platinum oxides was con— firmed earlier by Anson and Lingane (1957). The PtO and PtO2 are in 6 to 1 molar ratio regardless of degree of oxida- tion. Feldberg, Enke and Bricker (1965) defined three surface states of platinum electrode: reduced or "clean" state (Pt), the oxidized state (Pt(O)X) and the half-reduced state (Pt(OH)X). The Pt(OH)X film enhances electron transfer in the oxidation-reduction reaction. -A platinum electrode can be activated by strong oxidation and subsequent reduction producing the half-reduced oxide. Sawyer and Interrante found that reduction of oxygen at all preoxidized metal electrodes, except Ag and Au of all the various metals tested, appears to occur by the same mechanism. For the Ag and Au electrode, the oxide is reduced prior to the oxygen wave, and thus these electrodes act as reduced electrodes for oxygen wave. Mancy et al. (1962) in- dicated that better performance stability was achieved by using silver than using platinum. Wu (1967) showed silver produces lower residual current than platinum in oxygen concentration measurement. 10 The pH dependence of oxygen reduction at a platinum cathode was first shown by Laitinen and Kolthoff (1941). Their current-voltage curves indicate the reaction rate is quite independent of pH (within the tested range of 5 to 12) as long as the potential (vs. AgCl) is above -0.5 v, but varies greatly with pH as voltage becomes more negative. Unfortunately, they did not state the oxidation condition of the platinum electrode. It was made clearer by Sawyer and Day (1965) that oxygen reduction at oxidized electrodes occurs by a pH-dependent mechanism, whereas at reduced elec- trodes in acidic solutions oxygen is reduced by a pH-indepen— dent mechanism. Since platinum electrodes with different oxidized con- ditions may perform differently in the soil oxygen diffusion rate (O.D.R.) measurement, Black and Buchanan (1966) deter- mined the current-voltage (i-v) curve at pH 5.6 and 6.2. They found the i—v curves to vary with pH using oxidized electrodes but not with reduced and aged-oxidized electrodes. However, Black and West (1969) demonstrated the pH dependence of i-v curves in acidic soil, and attributed it to the re- duction of the carboxylic acid group of humic acid to the hydroxyl group at its solid state. The proposal is based on the fact that the i—v curve differs when the electrode is in a humic acid-water slurry from what it is in the supernatant liquid from the slurry. A similar curve was obtained when the electrode was in the supernatant liquid from a carboxylic 11 cation resin. The current was independent of pH at an applied voltage of -0.4 v (vs. S.C.E.). It is hard to evaluate their work because of their failure to mention the electrode conditions. III. THEORETICAL ANALYSIS In a polarographic determination of oxygen, the diffu— sion of oxygen to a metal wire anode follows Fick's second law in cylindrical coordinates. In an isotropical medium the law states ac _ 62c ;_a_c_ 1 52c 52c sz-DIa—r‘z+rar+;Za—ez+a—zs)' <1) where C is the concentration of oxygen at the point p(r,8,z) and time t, and D is the diffusion coefficient of oxygen in the medium surrounding the electrode. By assuming azimuthal symmetry of the cOncentration and if the exposed part of the wire is long, equation (1) reduces to a_c_ 62¢ liq -—-= O. (5) The general solution of (5) is C = A + B In r, where A and B are constants. Assigning the boundary conditions 12 15 C = C; at r a (radius of metal wire) C = C2 at r = b, a g_r g_b, the specific solution becomes (Ci - C2) ln ar C = C1 ln(a/b) ' Differentiating with respect to r gives d_c= 02-0: dr r ln(b/a) . (4) At time t > 0, C1, the concentration at the electrode sur- face, is much smaller than C2, the oxygen concentration at (b — a) from the electrode surface. .The flux written as f = D dC/dr at the electrode surface is then _ D C2 fr = a _ a ln(b/a) ' t > O ' (5) The flux is also related to the measured current by I=nFAf, (6) where I is the current in amperes, n is the number of elec- trons used for reducing one molecule of oxygen, F is the Faraday constant, A is the area of electrode in cm2 and the flux is in number of moles of oxygen transferred per second per cma. Substituting (5) into (6) gives _nFADCj I- a 1 n(b/a) ° ‘7) 14 If the diffusion coefficients of a set of media are all the same, and b is adequately chosen such that when an apparent steady-state2 is reached before reading the current, then the current passing through an electrode of fixed size is a func- tion of oxygen concentration C2 only. New let CQ=C2+€ where Co is the initial uniform concentration and e is any positive value. The value of e can be made as small as we want if we choose b sufficiently large. A precise treatment can be shown by relating it with equation (11). If b is adequately chosen so that e is negligible, then equation (7) becomes = n F A D C I _——Qa ln(b/a) (8) or I = k Co , (8a) where k is a constant and equivalent to the right-hand side of equation (8) except Co. Equation (8a) allows us to measure the oxygen concentration of a set of samples of identical diffusion coefficients provided one of the sample's oxygen concentrations is known. 2nAt t =¢§, the solution of equation (2), _ z_1_y _r ... C - szoY e dY where Z ‘ §_———n goes to zero. This implies a real steady'state can never~ e reached. -An apparent steady- state can, however, be arbitrarily chosen from the current- time curve where the slope is small without essentially losing validity of the steady-state assumption. 15 The accuracy of the measurement of oxygen concentration using a metal wire electrode depends on how close these assumptions are followed. Difficulties may be encountered when the foregoing theory is applied to measure soil oxygen concentration. For convenience the assumptions of equation (8) are iterated: (i) The diffusion coefficient is the same everywhere in the vicinity of the electrode. (ii) The diffusion is with azimuthal symmetry. (iii) The exposed part of the electrode is long. (iv) An apparent steady state must be reached before recording the current. (v) The value of b is chosen sufficiently large so that the difference between Co and C2 is negligible. (vi) The diffusion coefficient of all media of interest must be identical. This includes samples of known oxygen concentrations. For easy preparation of the latter, an aqueous salt solution is preferable. Assumption (1) can be satisfied if the medium of interest is isotropical, e.g., a liquid or a suSpension. Soil is hardly isotropical and particularly not as is seen by a thin wire electrode. When (i) is satisfied, (ii) will be satis- fied automatically for a wire electrode. ~Assumption (iii) re- quires a relatively long wire with respect to its radius. For the assumptions (iv) and (v), the value of b depends on the time chosen to reach a steady-state. A longer time gives 16 a better approach to a steady-state, but requires a larger b. A detailed account on these two parameters will be dis- cussed in the next section. Assumption (vi) is as difficult to fulfill as (i) in the case of soil. It is obvious now that the difficulty in determining soil oxygen concentration with a metal wire electrode arises primarily from the heterogeneity of soil to satisfy assump- tion (1) and (vi). 7A promising solution to the problem is to coat the electrode with a layer of porous material. If the thickness of the coating can be made equal to the parameter b required to satisfy equation (8) and if the coatings can be made so reproducibly porous that their diffusion coeffi- cients are identical, then assumptions (1) and (vi) are satisfied. .A minor deviation to satisfy assumptions (1) through (iv) may cause equation (8a) to be non-linear and to not intercepttfluaorigin in a current-concentration plot. The resulting errors in estimating the oxygen concentration may be corrected empirically. A common way to do this is to make a standard current-oxygen concentration curve and de- termine the oxygen concentration by interpolation. The following procedure could be used in the measurement of oxygen concentration. Let the coated electrode first equilibrate in the sample, which may either be an aqueous solution or a soil, then apply a negative current to the electrode of such a duration that the concentration gradient that is built up is essentially only within the coating. 17 At the end of this time a current reading is taken. A set of solutions of known oxygen concentrations may be used to plot the current-concentration relation. Oxygen concentra- tions of the unknowns can be found from the plotted curve of standard solutions. IV. THE POROUS COATING From the theoretical consideration, the porous coating on a thin wire electrode is eXpected to satisfy the follow- ing requirements: (1) The coating must be lyophilic and of adequate porosity. (2) It must be sufficiently thick. (5) It must be reasonably strong and rigid. The porosity of the coating should be such that the oxygen molecules as well as water and the principle charge- carrying ions can move freely through and within the coating. This can be determined eXperimentally by (i) the time re- quired to establish oxygen concentration equilibrium between the coating and the outside medium, and (ii) the electrical conductance of the coating in a salt solution. The coating thickness must be greater than (b - a) so that the requirement set by equation (8) is satisfied. Details will follow. The coating must also be able to maintain its rigidity. No significant deformation should occur when the electrode is inserted into a soil. Durability is also required for repeated usage of an electrode. 18 19 DETERMINATION OF THE THICKNESS OF COATING For simplicity, adequate thickness can be estimated by use of the linear diffusion law given by Ficks. In essence, any thickness greater than the required value of (b - a) is satisfactory. .For a thin coating, the linear diffusion law approximates well and will give a slightly greater value than is actually needed. Within a uniformly porous coating, the diffusion coefficient of oxygen may be considered constant, then it follows where C is the oxygen concentration at a distance x from the electrode surface at time t, D is the diffusion coefficient. A solution of the equation is _ 2 C = a f e 2 dz, (9) where z = x , (10) ZVD t and a is a constant determined by boundary conditions. At the initial and boundary conditions, Co (initial concentration) when t = 0 C C = 0 at x = 0 when t > 0 , equation (9) becomes fz e“y dy . (11) 20 The right-hand side is called the error function and its values are tabulated versus 2 in many books, e.g., Kolthoff and Lingane (1952, page 25). If x is taken as the thickness of the coating, then C is the concentration at the outer boundary of the coating. It is desirable to have C as close to Co as possible so the oxygen gradient is built up essentially only within the coat- ing at time t during which the current is recorded. Under this condition, the diffusion coefficient D in equation (10) is determined by the coating only. This satisfies the requirement of constant D for equation (8). From the assigned value of C/Co, a corresponding value of 2 can be found. If the diffusion coefficient in the coat- ing is known and t is adequately chosen, then x, the thickness of the coating, can be calculated from equation (10). Here t is the time required to approach a steady-state. A longer time gives a better approach but requires a thicker coating. Therefore t is chosen at a value at which the thickness of the coating reaches its upper optimum limit. On the other hand, a thin coating is preferred, because a thick coating requires a longer time, prior to t = 0, to equilibrate the oxygen concentration inside and outside the coating and at the same time allow the surrounding medium to recover to its original oxygen concentration. In addition, a thick electrode, in which the oxygen concentration is to be measured, tends to disturb more of the soil than a thin electrode and is not desirable. 21 ESTIMATION OF THE DIFFUSION COEFFICIENT IN THE COATING In order to find 2 by equation (10), D must be known. Estimation of D in the coating is possible by applying equa- tion (8). Let I; and I2 denote the currents measured by using coated and uncoated electrodes, respectively. If the measurements are made in one solution and the electrodes have the same exposed areas, then by equation (8) we have nFADLQQ 11 = a ln(bl/E) °°ated = nFADg cg I2 a ln(b2/a) .uncoated, where D; is the diffusion coefficient in the coating and D2 is the diffusion coefficient in the solution. Usually D1 is smaller than D2 due to the torturosity of the diffusion path in the coating material. Hence b2 is usually greater than b1, when both satisfy equation (8) to the same extent or produce the same value of e, as can be seen from equation (10). In fact any increase in the thickness of the coating beyond b; gives a deviation of the estimated concentration insig- nificantly small and within tolerance of the instrument. .The value of b1, therefore, need not be as large as b2. Let us simply imagine the outer solution to be an additional coat and make the thickness of the entire coating equal (b2 - a). Dividing I; by I2 gives 22 £1. = 21. or D2 = D-J—a-I: . (12) Let D; be the diffusion coefficient of oxygen in water which is known, then D2 can be calculated from the two current readings. A STANDARDIZATION PROCESS It is difficult to make the collodion coatings reproduc- ibly porous so that the diffusion coefficients of oxygen in the coatings are all identical. In addition, the porosity of a coating may change with time, and so does the "reactive" surface area of the electrode. If these changes occur the current measured will be influenced according to equation (8). A simple but effective way to minimize these variations is to use the same electrode and take a current reading from a solution of known oxygen concentration (air saturated KCl solution is most convenient) after the measurement of an un- known. For the convenience of discussion, we define: (1) gtangardized current as the ratio of the current reading of the unknown sample to that of air saturated 0.02 N KCl (the two measurements should be made consecutively at the same temperature and atmospheric pressure using a single electrode): (2) oxygen saturation as the ratio of quantity of oxygen present in a solution to the greatest amount possible 25 at the same temperature and pressure. Complete saturation of the solution is designated by 100% oxygen saturation and that equilibrated with air as 20 (air has 20%onygen by volume). Denoting Ix and Cx the current and concentration, resPectively, of the unknown, and Ia. and C . that of the ir air air-saturated solution and substituting these values into equation (8a) yields ’f = x (15) The left-hand side of (15) is the standardized current of the unknown. Since Ca is a constant at a specific temperature ir and pressure, standardized current is, therefore, a function of the concentration of the unknown only. The dependence of the current on the surface area of the electrode and diffusion coefficient of oxygen in the coating is eliminated. It might be better to eXpress oxygen concentration in a soil in terms of oxygen saturation. This expression describes both the oxygen concentration in soil air and soil solution at electrode surface provided they are in equilibrium. The standardized current and the oxygen saturation are independent of temperature and atmospheric pressure at which measurements are made; hence, a current vs. concentration curve eXpressed in terms of these two parameters can be used as a standard at all temperatures and pressures. V. THE OXYGEN CONCENTRATION METER Applying the foregoing principles, an instrument was constructed for measuring soil oxygen concentration. A silver wire is used as the cathode and silver—silver chloride in saturated KCl solution as the anode. ‘After a short duration of circuit closure, a portion of transient current is inte- grated and amplified. The output signal of the integrated current is read as voltage which relates directly with oxygen concentration. Ten electrodes can be used at a time. Block and wiring diagrams are shown in Figure 1, and the wiring diagram and description of the panel are in Appendix A. The main parts of the system are discussed separately. For added convenience, the meter is also wired for taking oxygen diffusion rate measurements. THE ELECTRODE SYSTEM Although platinum has been commonly used as the cathode in polarographic measurement of diSSOlved oxygen, the require- ment of platinum oxide at electrode surface complicates the oxygen reduction mechanism (Lingane, 1961). On the other hand, the oxygen reduction mechanism is independent of the 24 25 .mcfipuoumu How uoumummuca mnu ou Hofiwu pcoomm we» nmsouzu cmau can wumumlmpmwum ucmummmm cm nmwanmumw 0» mafia meouum Uwaqm may nmsounp mommmm umnwm ucwuusv- .umuwe coflumuacmocoo commxo may no Emummfiv xoon .a musmwm HmumEuHo> _.I IIIIIIIIIIIIIIIII I. _ _ _ _ _ _ ..L _ Houmummucw HmEHu HmEHu ucmunso .IIIIII pcoomm III: pun?” >Hmmsm mam can Emumwm mpouuumam 26 previous history of oxidation at silver and gold surfaces (Sawyer and Interrante, 1961). By using a silver instead of platinum electrode in oxygen concentration measurement, better stability (Mancy et al., 1961) and lower residual current (Wu, 1967) were achieved. For these reasons, silver is chosen as the cathode material. The cell consists of a rectangular plastic box contain- ing a loosely folded silver £0113 in a saturated KCl solution with KCl crystals. The silver sheet is plated with AgCl. The box connects to a fritted filter candle through tygon tubing. The fritted section has a length of 2 cm and a diameter of 1.2 cm. To assure good electrical continuity with the soil sample, the bottom of the candle must also be porous. The fritted part is impregnated with 1.5% agar in saturated KCl. This slows down the mass flow rate of the saturated KCl solution into the sample without appreciable impairment to the conductance of the system. -A 6 cm long 22 gauge silver wire is used to construct the electrode. The wire tip is first heated until one end is just rounded without protrusion in diameter. This can be accomplished by heating the wire near the bottom of a small gas flame. The tip is further heated so that the 1 cm end section becomes dull. .This provides a better grip of the coatings to the electrode. The rounded tip is then coated 3It is the same one used in the commercial oxygen diffusion maters made in Dicks Machine Shop, Lansing, Mich. 27 with tygon paint (black, TP-21, U. S. Stoneware, Akron, Ohio 44509) to a length of 1 to 2 mm by withdrawing the wire slowly from the paint so that the coating is thin and uni- form. It was repeated twice; allowing for drying between coats. The other end of the wire is soldered to a plastic covered 22 gauge stranded c0pper wire. If the sites to be measured are more than 5 cm deep, the wire is first soldered to a plastic covered jumper wire then to a stranded wire. The exposed portion of the wire including the soldered con- nection is painted with tygon but a 4 i.0.5 mm exposure is left at the end. The soldered part is wrapped with electric tape and two additional coatings of paint are applied. If the electrodes are not to be used fairly soon, it is sug- gested the electrodes be stored at this stage rather than a later stage. The collodion solution is prepared by mixing 10 parts of collodion (Mallinckrodt, U.S.P.) with 1 part (by weight) of glycerol. Since the collodion solution is extremely vola- tile, the following provides a quick way of preparation of the mixture without excessive evaporation. 15.1 5 g of collodion is poured into a tared narrow-necked 50 ml bottle, stoppered and weighed on a single-pan balance. While keeping the bottle on the balance, the required amount of glycerol is added with a medicine dropper. The contents in the bottle are mixed well in a reciprocating shaker. To achieve a better reproducibility of coating thickness and porosity, the 28 mixture is preferably prepared just prior to the application of the coating. Before applying collodion coating, the tygon-coated wire is washed by rubbing it between gloved fingers with floride- free toothpaste. Cleaning was stopped before the silver surface becomes shiny again. The wire is rinsed thoroughly in tap water and then placed in distilled water. After all electrodes are cleaned, they are dried with tissue paper and air dried for a few minutes. To coat the electrodes, they are dipped directly into the bottle of collodion mixture, then air dried for 5 seconds. While drying the electrodes are held horizontally and kept turning. .They are held only by the plastic covered copper wire, not the silver part (which is an excellent heat conductor) to avoid evaporation of the solvent from inside the coating. Electrodes with visible bubbles within the coating should be discarded. The coating process is repeated 8 times, then the electrodes are air dried for 60 min. The electrodes are stored in a 250 ml flask half—filled with 0.02 N KCl. It is important that the coated electrode not be taken out of the KCl solution and exposed to air for more than a few seconds. Longer exposure may cause sufficient drying to ruin the coating. If longer exposure is desired, such as in the case of checking the thickness of the coating, the electrode should be dipped in 1:1 glycerol- water solution before exposure. The collodion mixture must be kept stoppered to avoid evaporation and should be swirled 29 well, which produces less foam than shaking, before coating each electrode. A new electrode usually has a high resistance which can be lowered by passing a current at —1.2 v (vs. AgCl) for 15 minutes. A good electrode should give a resistance of less than 4 kilo-ohms between the AgCl and the electrode, while in 0.02 N KCl. THE TIMING RELAYS Two identical time-delay relays‘ are used in series. The activation of the first relay is simultaneously accom- panied by a closing of the circuit of the electrode system. At the end of 0.4 second, the first relay activates the second relay which sends the current to the integrator for 0.1 second. THE INTEGRATOR The transient current flowing through the electrode is registered by the Keithley5 model 501 solid state electrometer operational amplifier. The amplifier integrates the input current I and yields an output voltage V as given by ‘Intermatic model $810225B-1, International Register Company, 4728 W. Montrose Ave., Chicago, Ill. 60641. The model is a surface mount type and has an operating range of 0.05 to 0.5 second. 5Keithley Instruments, Inc., 28775 Aurora Road, Cleveland, Ohio 44159. 50 f1 dt . (14) H: The degree of amplification is determined by the feedback capacitor Cf. At steady state,6 I is independent of t, then _ I V--E—At. (15) f If At, the duration of time in which the current is integrated, is invariant, and the same feedback capacitor is used in all measurements, then the measured voltage V is directly propor- tional to the input current I. Applying equation (15) to (15) yields . (16) air air air where Vx is the voltage measured from the unknown and Vair is the voltage measured from the air saturated solution. The voltage ratio is, therefore, the standardized current. THE VOLTMETER A 20 milliampere ammeter is used to construct the volt- meter. It gives a full-scale reading of 1.0 v, 0.5 v and 6If the steady-state is not obtained but At is small, the current-time curve withinIAtcan be approximated by a straight lineA The average I ‘withinzxtis then simply the current at t + §£-. We can substitute this average :[ for I in equation (14) and also obtain equation (15). 51 0.25 v for measuring the output voltage from the integrator. The meter serves also as the indicator of the O.D.R. current, the emf of the electrode system, and for checking the voltages of the batteries which power the integrator and timers. VI. RESULTS AND DISCUSSION Bare electrodes were used in the early stage of this study to measure oxygen concentrations in bentonite sus- pensions and dilute potassium chloride solutions. Highly reproducible current readings were achieved under carefully controlled surface area and surface conditions of the elec- trodes. It was found effective to rub electrodes with tooth- paste in providing uniform surface conditions among electrodes. .Measured current-oxygen concentration relation— ship agreed well with that given by Brandt (1965) and Wu (1967). However, when measurements were made in sand and silt, readings became erratic and were usually lower than that measured in pure solution of the same oxygen concentra- tion. Theoretical analysis of the problem, as discussed in section III, pointed out that oxygen concentration is pro- portional to current only when the diffusion coefficients of oxygen in the sample media are constant. This constancy may not be obtained in soil due to its diverse combination of solid, liquid and gaseous phases. Coating the electrode with a porous medium was then pr0posed and a search for such porous medium was conducted. It was found that electrode coated with 10:1 collodion-glycerol mixture provided a good 52 55 porous coating. The coating is easy to prepare and adheres to the electrode very well and has sufficient strength and rigidity. Experiments were then conducted to test the adequacy of the coating thickness as well as the factors that may affect the oxygen concentration measurement. In measuring oxygen concentrations, the electrodes and the salt bridge were inserted into the sample 1 cm above the site to be measured; while in aqueous solution, the electrodes were inserted directly into the final positions. An emf of —0.65 v was applied to the electrodes for five minutes. .The electrodes were then inserted 1 cm deeper and after a 5 minute wait, to let the oxygen concentration in the sample equilibrate with that in the coatings, the circuit was closed and current readings were taken. .A detailed pro- cedure is shown in Appendix B- THE COLLODION COATING Calculation of coating thickness, by using equation (11), requires the knowledge of (i) the concentration ratio C/Co, where C is the oxygen concentration at the outer boundary of the coating and Co is the initial oxygen concentration, (ii) the diffusion coefficient of oxygen in the coating, and (iii) the time required to reach an apparent steady- state. .The choice of the C/Co in (1) determines the accuracy of the instrument. For the second parameter, a value close 54 to the upper limit will minimize the possibility of under— estimating the coating thickness. The third has little effect on the accuracy of the measurement because by using standardized current the errors will cancel out. However, it is related to the sensitivity of the measurement as will be seen later. The concentration ratio The best choice of the concentration ratio should give a deviation similar to the component of the instrument that gives the poorest reproducibility. The timers which have a reproducibility in the order of 1%imost likely have the poorest reproducibility of all the components. .The C/Co is, therefore, set at 0.995. This restricts a concentration drop of only 0.5% at the boundary provided the diffusion coefficients of oxygen in the soil and in the collodion are the same. .Any moderate departure of the two diffusion co- efficients may not change the concentration at the boundary very much and the ratio C/Co could be considered as un- altered. This ratio gives a 2 value of 2.0 by equation (11). The diffusion coefficient in the coating The oxygen diffusion coefficient in the collodion coat- ing depends greatly on the porosity and age of the coating. The porosity can be controlled somewhat by the ratio of glycerine to collodion. .More glycerine gives a higher 55 porosity. By applying equation (12), the estimated diffu- sion coefficient ranged from 1/2 to 2/5 that of water. In water D has a value of 2.58 x 10‘5 cm2 sec-1 at 25°C (Millington, 1955), and 2/5 of it yields a'D value of 1.6 x 10"5 cm2 sec'l. The choice ofrthe steadyrstate time Since there is no real steady—state in the oxygen dif- fusion current the choice of steady-state is rather subjec- tive. However, the current-time relationship, as revealed by the oscilloscope photo (Figure 2), indicated that the background current (the N2 curve) decreases, relative to that of air saturated solution, with time. This means the sensitivity of the instrument can be improved by increasing the time. The improvement is greatest before 0.5 second and relatively slow thereafter. In order to have a useable sensitivity and have a thin coating the t in equation (10) was chosen as 0.45 second and the current was integrated, between 0.4 and 0.5 second. These values substituted into equation (10) yield a coating thickness of 0.107 mm, which is about 1/5 the radius of the silver wire. A test was made to show whether the calculated thickness is appropriate. Testing the thickness of coating A preliminary examination was made to compare the vari- ability of measured oxygen concentration of a several Current Figure 2. 56 0.0 0.2 0.4 0.6 0.8 Time (seconds) Oscilloscope photo showing current-time relationships after a -0.65 v emf was applied to a collodion coated silver electrode. The upper curve was obtained from air- saturated 0.02 N KCl solution and the lower curve from N2 saturated. The base line for both curves is the hori- zontal line just below the lower curve. 57 thicknesses of coating at different conditions. Current readings were taken in medium sand and silt samples in 0.02 N KCl at various moisture tensions. It seems evident, as shown by Table 1, that readings from thin coatings were affected more by the blocking effect of solid particles, moisture tension and consequently the presence of air than the thicker coatings, whereas the bare electrodes showed the highest variation. The same electrodes were then tested by the variances (Table 2) at the most discriminating conditions, i.e., using the silt sample at 70 cm tension. In each of the five coat- ing thickness readings were obtained from 2 electrodes of similar thickness with 11 repetitions. The estimate of the component of variance due to measurements is over 15 times as large as the component due to electrodes. This means that the variance of a single observation on an arbitrary electrode 52 3M2 + 532 = 0.00147 + 0.0000881 ,0.00156 has much smaller error due to electrode construction than due to the error of measurement. .This single observation gives a standard error i.4% of the standardized current, or less than i.1.5% of the Oxygen saturation. The latter may drop to 1.0.5%$when single measurements are made using 10 electrodes to determine the oxygen concentration in a soil. 58 .uawm &a0 0cm wean wean >nm> &00 mcflmucoo uaflm 0:9 .930 0:3 &00 can 05mm 89.30.: &0> mcflmucoo 050m 2:509: man. 00.H 00.0 H0.d. no.6 00.H 00.0 ddd.0 0$dlom 00.H 00.6 00.d N0.d 00.H 00.0 ddd.0 m$dlom 00.H 00.0 d0.fi 00.0 $0.d 00.d 0>0.0 $$dI0N 00.d 00.0 $0.H 00.0 00.a 00.0 000.0 0$dlom 00.H 05.0 $0.a >0.0 00.0 H0.H «00.0 N$HION 00.6 $5.0 00.H 50.0. 00.0 H0.d H00.0 d$d|0N 00.fi «0.6 $0.6 00.0 $N.H ma.d 000.0 0NHION 00.d >0.0 00.H 00.0 mm.d Nd.a 000.0 SNHION 00.H $0.0 00.0 00.0 00.0 00.0 0 N10 00.d $0.0 $0.0 «0.0 00.0 «0.0 0 dl0 cowmcwu coflu cofimcmu cofimcmu code 80 0h Imusumm EU 00 EU 0$ Imusumm mace EB Hum aux 2 No.0 CH HUM 2 No.0 CH mmmcxownu H0985: 2 No.0 uaam wean Eswcmz mewumoo woouuomam ucmunsu vmuflcnmvcmum .mmmmmcxuwnu mcflumou msoHHm> mo mmwonuowaw an UNHDWMOE mm MHMvOE UGHMHDUNM HHM EOHH MHCOHHDU @ONHUHMUGMUG Gnu. MO COWHHMQEOU .H OHQMB Table 2. 59 Analysis of variance for the standardized currents measured in a silt sample leached with air satu- rated 0.02 N KCl and equilibrated at 70 cm tension using electrodes of various coating thicknesses. Source of variation d.f. M_S Expected Mean Square . 2 2 2 Thickness 4 1.042 6M + 116E + 22KT Electrode 5 0.00244 0M2 + 116E2 Measurement 100 0.00147 6M2 With m measurements per electrode and e electrodes the variance of the mean of and when m si X Thus Tukey's critical value 0.05 As shown by Table = 11 and e 5 thickness is 2 0.000101. is QO.OSS§-= (5.67)(0.0100) 0.0567 the differences between the coated and the uncoated electrodes were significant at 5% level, but the differences between the coated electrodes were non- significant. In other words, coating thickness as thin as 40 Table 5. Differences of the standardized currents from electrodes of various coating thicknesses as measured in a silt sample leached with air- saturated 0.02 N KCl and equilibrated at 70 cm tension. Coating _ _ _ _ _ _ _ _ _ thickness xi x144 x0 x1 - x1 x1 - x5 xi - x4 I— ——————— I- ———————————————— x2 0.051 1.0086 0.4957 : 0.0245 0.0055 .0.0005 I x4 0.111 1.0081 0.4952 : 0.0258 0.0050 I x:5 0.079 1.0051 0.4882 I 0.0188 I x1 0.056 0.9485 0.4694 I II ------ x0 0 0.5149 E; The differences above the dotted line exceed the critical value 0.0567 at 5% level for Tukey's test. 0.056 mm, which is only 1/5 that of calculated thickness, seemed satisfactory and an increase in the thickness gave little improvement in the consistency of the standardized currents. It should be noted, however, that the adequacy of the coating thickness depends on the diffusion coefficient in the testing sample. .Media denser than this silt (which is quite dense) could increase the current differences between 0.056 mm and the thicker coatings. Therefore, the calculated thickness which provides a safety factor of over 5 with respect to the tested condition, may be reasonable. The adequacy of the coating thickness may also be revealed by the standard curves of various coating thicknesses. 41 The coating thickness and sensitivity A standard current-oxygen concentration curve is re- quired to convert standardized current into oxygen satura- tion. Figure 5 shows the curves made by using electrodes of various coating thicknesses. The curves indicate that oxygen concentration resolution decreased with increasing coating thickness but approach a limit at a thickness some- where between 0.111 and 0.079 mm. This reveals that the standard curve is dependent on the coating thickness when it is insufficient; and it is independent of coating thick— ness when it is near the calculated thickness, i.e., 0.107 mm. This supports the calculated thickness as a good approximation. The top curve in Figure 5 may be used as the standard curve for converting standardized current into oxygen saturation. Standardized current 42 1.00 0.80 0.60 Coating thickness: 0 0.111 mm A 0.079 mm 0.20 O 0.051 mm A 0.056 mm C] 0 (bare) O I I I I L I l I I J I i I I J I I l I J 0 5 10 15 20 Oxygen saturation, % Figure 5. Standard curves obtained from electrodes of various coating thicknesses. 43 THE4E§FECT OF APPLIED E.M.F. A flat plateau in an ideal polarographic curve would indicate that the current is independent of the applied voltage in the vicinity of the chosen emf for measurement. -Figure 4 shows the emf and current relationship of a collodion- coated electrode. The gentle slope between -0.6 and -0.8 volt indicated that the current depended only slightly on emf. ~An emf of -0.65 volt was chosen for oxygen concentra- tion measurements. Current, meter readings o I I I I I I I l J -0.2 -0.4 -0.6 -O.8 -1.0 E applied, vs AgCl electrode, volts Figure 4. Oxygen diffusion current at various applied emf. 44 THE EFFECT OF SALT CONCENTRATION The oxygen reduction current is conducted by ions in the solution of the sample. In soil, most of these ions are in the form of dissociated salt. The overall concentra- tion of these ions should always be sufficient so that the current is limited by oxygen concentration only. .It is necessary to find the lower limit of salt concentration at which the measured oxygen concentration is within acceptable precision. To Specify this limit, currents were determined in air saturated solutions of various KCl concentrations. Within a wide concentration range of 0.0015 to 0.05 N KCl, a 2.4% variation of standardized current was obtained (Figure 5). This corresponds to.i 1.5% of oxygen satura- tion. The current increased slowly with decreased salt concentration and reached a maximum at about 0.005 N KCl, then fell down slowly. ~An interesting phenomenon was that the maximum coincided with the current-resistance curve (Figure 6) where the resistance was controlled by an added variable resistor in series with the cell system. .Both maxima had a total cell resistance of about 5 kilo-ohms. It was concluded then that the change of the standardized-current with salt concentration accounted mainly for the resulting resistance effect. A possible explanation of the increased current with resistance is that the increased resistance may restrict, as expected by Ohm's law, the current flow in the smuo-otrx 'meisKs eporiaete sq; go aoueistsez teiom .Gowumusu0m new um mcoflusaom HUM cw Umusmmmfi mucmnnso coflmsmmwo cmmmxo coflusaom Hux.mo cofiumnucmucoo undo: .0 musmflm $00.0 00.0 No.0 «0.0 000.0 N00.0 N _ A _ _ _ fi 4 d ONoo T I $ T I.0$.0 mocmumflmmm I , I 0 I IIO0.O l 1 0 T .100.0 CHAHI I 100.9. ucmHHDU O NHI Low. a quezzno pezthEpueqs 45 46 added resistances. St . andardlze Cu 1.00! rrent _10 O O *8 a) 0.80— . _ 8 n H 5 e O C S 0&3? N 0.60— ‘39 . —I 6 :3 05» H ¢° m D "O a 4‘3 0.40— . a 4 m 0.20.. _ 2 I I l I I 0 1 2 5 4 5 6 Added resistance, kilo-ohms Figure 6. Oxygen diffusion currents measured at various Total resistance of the electrode system, kilo-ohms 47 very early stage of current flow where much higher current drain is required (see Figure 2). The formation of a concen- tration gradient may thus be slightly retarded even at later stages when the current is not limited by salt concentration. The gross oxygen concentration is, therefore, slightly higher than would be expected, and results in a corresponding higher current reading. The coincidence of the two curves of Figure 5 and 6 suggests such deviation can be corrected, at least to some extent, by the following method. The system resistance of the unknown is measured following the current measurement. The resistance of the air saturated 0.02 N KCl system is then Vadjusted to the resistance of the preceding sample with a variable resistor in series before the reference current read- ing is taken. The ratio of the two readings gives a resistance- corrected standardized current. Since the standardized cur- rent drOpped rapidly beyond 8 kilo-ohms in Figure 5 but did not show so in Figure 6, oxygen concentation measurement will thus be inaccurate in media with resistances higher than 8 kilo-ohms. The resistance can be conveniently measured by 7 Since error due to salt con- the BOuyoucos moisture meter. centration accounts for less than 2% of oxygen saturation, the correction for resistance might not be necessary when 7Model BN-2A, Industrial Instrument Inc. .The portable unit produces an alternating current for resistance measure- ment and has a full—scale resistance of 10 kilo-ohms. 48 that error is acceptable; but such a resistance measurement is still strongly suggested especially when the sample's resistance is low.8 OXYGEN CONCENTRATION MEASUREMENTS IN FLOODED AND WE! SOILS The instrument should be useful in monitoring the oxygen concentrations in paddy soils or wet soils. In such situa- tions oxygen depletion is likely to occur. To test the instrument, a number of flooded conditions were simulated in the laboratory and greenhouse in which oxygen concentra- tions were measured. Measurements were also made in the field when the soil was wet. The results are discussed as follows. Two experiments were made in flooded soils. In the first eXperiment two pots were filled with air dried loam soil. One pot was watered with aerated tap water and the other was watered with non-aerated tap water. Oxygen concen- trations were measured right after the watering. Six to 10 readings were taken from each pot. The average oxygen saturation of the soil irrigated with aerated tap water was 11.1.: 1.5%9 and that of the soil irrigated with non-aerated 8For instance, the electrode itself may change its con- ductance due to excessive exposure to air. 9Standard error for the average value. 49 tap water was 2.2 i 0.4%. Measurements were repeated two days later, after watering, and the average oxygen satura- tions of the soil irrigated with aerated and non-aerated tap water were 8.9 i.1.5% and 2.0.1 0.7%, respectively. The results indicated that non-aerated tap water has a very low oxygen concentration and irrigation with such water could be detected with the instrument. In the other experiment, two soil samples of Sims sandy loam were taken from a rotation experiment at the Ferden Farm, near Chesaning, Michigan. Rotation 6 consisted of corn-sugar beet—barley-bean-wheat, whereas rotation 1 con— sisted of corn-sugar beet-barley22 years of alfalfa brome. Rotation 1 showed a better long-term average physical con- dition than rotation 6. To assure soil homogeneity, so that consecutive concentrations could be comparable, soil was first air dried and passed through a 2 mm sieve. A plastic tray (27 cm x 19 cm x 6 cm deep) was half filled with air saturated 0.02 N KCl. Soil was poured slowly into the tray until the soil was 4.5 cm thick, but with the KCl solution constantly covering it. More KCl solution was added until it was 0.5 to 1 cm above the soil surface. The tray was covered with another tray but with a small gap for air exchange without excessive evaporation. The temperature was maintained at 25 1.1OC. Aerated distilled water was added following measurement to replenish evaporation loss. The results are shown in-Figure 7 where each point was the 50 news: meumensm mcflmn “mumm maflom 00>mflm cw coflumuucmucou cmmmxo 0 fl .HOM 2 No.0 vmumHDDMmIHflm mewooon Hmumm 0000. 0 N H 0 mommnsm momwusm mommusm mommusm Meow snow show Hsom 30Hmn EU $ 3oH0£ umsn BonQ EU $ 30Hmn umSn fl \ s s s s 0 0 a GOHDMDOH H E cofiumuou coflumuou coflumuou 00H mvcmm mafia . 00 .mommusm um 0:00:0um uwumz .00Nfla Ifluumu 0:0 >00 msofl>mum 00p msofi>0um onu A00 00 Ahv 00 CH ca0n >>0wm ca cfimu >>00m_ 0800 0E00 coflummfiuufl HmecHHmm xumfiwm 00.0 00.« 00.0 00.0 00.0 00.0 $0.0 N0.o m\\m .Houum II 0H00c0u0 0$.N fi0.0 00.H 0H.N 00.N N0.N a>.a 00.fi m .coflu0fl>00 0H00c0u0 00.N N0.$ 0H.N 00.N H0.$H 00.0 0 NH.0 0 0>.N 000H0>< III 0.0 $.$ H.N 0.0a mnmfi o 0.0 III III 0.0 0.0 0.H $.0 0.0a $.> o N.N 0 0.H 0.0 0.a 0.0 0.H N.>H H.0d o d.0 0 N.$ 0.0 0.0 N.d $.0 0.NH H.0 0 0.0 0 0.0 0.0 0.a $.0 $.0 0.0a 0.0 0 0.0 0 N.H $.0 0.0 0.N fi.0 H.Na 0.0a 0 N.0 0 H.a 0.H 0.0 0.0 0.0 0.$fi 0.0 0 0.0 o 0.0 0.0 0.fi 0.d >.0 0.$fi N.0 0 0.> 0 0.a >.a 0.0 $.m 0.0 0.$H 0.0 0 m.$ 0 H.0 0.fi 0.0a 0.H 0.0 0.00 N.0 0 0.0 o >.0 0.0.0 R coHu0H5u00 :0 ucwmmmmmmwfl Muse _ 000m MADE _ wumm muse fl. 0H0m_ muse _ mumm Hm>oo mono m>00 0:00000HQ 03v :0 n>00 msoH>0Hm coflufl0coo >cflmn pan >00 wnu :0 H000 0cm >ccsm msu 020 >00 00p :0 >cc50 H0£u003 Ifi. 000.0H 000N 000m .ME0B HHom Osma .m mesh mm >0: am 02 0009 _._._. ._. NH. 0:» D0 A80 $INV Hfiom QOD msu GA mmumu :oflmSMMH0 cmm>xo 0cm mcoflumuucmucou com>xo .Eumm mHflom m.>uflmum>wcb .e 033 55 The same trend was found for O.D.R's. The high oxygen con- centration in the turf-covered soil may be due to the cool rainy weather in the preceding days, and the loose matrix at the soil surface and the incomplete watering prior to take measurements. The fallowed plot had water standing while the turf did not. The thin wire electrodes had no dif— ficulty to be inserted into the saturated bare soil but did find difficulty at times (about 2 out of 10 electrodes) when inserting them into the densely rooted turf-covered soil. In the latter case, about 50% of the electrodes were with broken coatings after couple times of service but none in the former case. Readings from broken electrodes were excluded from Table 4. VIII. SUMMARY AND CONCLUSIONS An instrument that is capable of measuring oxygen con- centration in soil at or near saturation has been designed, constructed and tested. It uses a collodion-coated, silver electrode less than 1 mm in diameter. With an electrode of this size good spatial resolution can be achieved without seriously disturbing the soil structure. The instrument measures an integrated current 0.4 second after closing the circuit. The reading is independent of oxygen diffusion coefficient of the soil. ~Drift due to change of surface character of electrode is minimized by taking a reading in air saturated 0.02 N KCl immediately following the unknown, both are at the same temperature. The ratio between the two currents, called standardized current, is a measure of the oxygen concentration of the soil. Oxygen concentration expressed in terms of oxygen saturation eliminates the re- quirement for temperature correction. The resistance of the cell system and soil influences the standardized current. The effect of resistance is corrected by putting a variable resistor in the system so the current reading in air saturated KCl solution can be measured at the same resistance as that of the preceding unknown. Oxygen concentration 54 55 measurement in media with resistances between the Ag elec- trode and the reference cell higher than 8 kilo-ohms are not suggested because such high resistances are not correct- able. The instrument has been tested in saturated soils in the laboratory, greenhouse and field. It should be useful in measuring oxygen concentration in paddy soils and in other soils following irrigation or a heavy rainfall when oxygen depletion is likely to occur. BIBLIOGRAPHY BIBLIOGRAPHY Anson, F. C., and J. J. Lingane. 1957. Chemical evidence for oxide films on platinum electrometric electrodes. J. Am. Chem. Soc. 79:4901-4904. Bergman, H. F. 1959. Oxygen deficiency as a cause of disease in plant. Bot. Rev. 25:417-485. Black, C. A. 1968. Soil-plant relationships, John Wiley and Sons, New York, pp. 155-201. Black, J. D..F., and A. S. Buchanan. 1966. Polarographic reduction of oxygen at a platinum surface in relation to the measurement of oxygen flux in soils. 'Aust. J. Chem.-19:2169-2174. Black, J. D. F., and D. W. West. 1969. Solid-state reduc- tion at a platinum micro-electrode in relation to measurement of oxygen flux in soil. >Aust. J. Soil Res. 7:67-72. Blinks, L. R. and R. K. Skow. 1958. The time course of photosynthesis as shown by a rapid electrode method for oxygen. Proc. Natl. Acad. Sci. 24:420-427. Brandt, G. H. .1965. The estimation of dissolved oxygen concentrations in soils with bare, stationary platinum wire electrodes. Ph. D. Thesis. .Michigan State University. Carritt, D. E. and J. W. Kanwisher. .1959. An electrode system for measuring dissolved oxygen. ~Anal. Chem. 51:5-9. -Ciufk3uLs C., R. Wolf, D. Granger, and Z. Taylor. .1955. Continuous recording of blood oxygen tension by polarography. J. Appl. Physoil. 6:189-195. Clark, L. C. 1956. Monitor and control of blood and tissue oxygen tensions. Trans. Am. Soc. Artificial Internal Organs. 2:41-48. 56 57 Danneel, H. 1897. Uber den durch diffundierende Gas hervorgerufenen Reststrom. Zeits. f. Elektrochemie 4:227-252. Davies, P. W. and F. Brink. 1942. Microelectrodes for measuring local oxygen tension in animal tissues. Rev. Sci. Instruments-15:524-555. Drenckhahn, F. O. 1951. Untersuchungen zur polarometrischen Messung des Sauerstoffdruckes (p02) in Blut mit der Platinelektrode. .Naturwissenschaften 58:455-456. Enoch, H., and V. Falkenflug. .1968. An improved membrane system for oxygen probes. Soil Sci. Soc. Amer. Proc. 52:445—446. Erickson, A. E., J. M. Fulton, and G. H. Brandt. .1964. New techniques for relating soil aeration and plant response. Trans. 8th International Congress Soil Sci. 1:171-176. Feldberg, S. W., C. G. Enke, and C. E. Bricker. 1965. Formation and dissolution of platinum oxide film: mechanism and kinetics. »J. Electrochem. Soc. 110:826-854. Flynn, D. S., D. G. Kilburn,.M. D. Lilly, and F. C. Webb. 1967. .Modifications to the Mackereth oxygen electrode. Biotechnol. Bioeng. 6:625-625. Grable, A. R. 1966. Soil aeration and plant growth. Adv. Agron. 18:57-106. Hack, H. R. B. 1956. An application of a method of gas microanalysis to the study of soil air. Soil Sci. 82:217-251. ' Hutchins, L. M. 1921. Oxygen supplying power of the soil. Plant Phys. 1:95-100. Kolthoff, I.-M., and J. J..Lingane. 1952. Polarography. 2nd ed. Vol. 1, p. 25, Vol. 2, pp. 411-418, 552-558. Krog, J., and K. Johansen. 1959. Construction and charac- teristics of teflon-covered polarographic electrode for intravascular oxygen determination. -Rev. Sci. Inst. 50:108-109. Laitinen, H. A., and I. M. Kolthoff. 1941. Vbltametry with stationary microelectrodes of platinum wire. J.-Physi- cal Chem. 45:1061-1079. 58 Lemon, E. R., and A. E. Erickson. 1952. The measurement of oxygen diffusion in soil with a platinum microelectrode. Soil Sci. Soc. Am. Proc. 16:160-165. Lingane, J. J. 1961. Chronopotentiometric study of oxygen reduction at a platinum wire cathode. J. Electroanal. Chem. 2:296-509. Mancy, K. H., D. A. Okun, and C. N. Reilley. 1962. A gal- vanic cell oxygen analyzer. J. Electroanal. Chem. 4:65-92. McIntyre, D. S., and J. R. Philip. 1964. A field method of measurement of gas diffusion into soil. Aust. J. Soil Res. 2:155-145. Millington, R. J. 1955. Diffusion constant and diffusion coefficient. Science. 122:1090-1091. Raney, W. A. 1950. Field measurement of oxygen diffusion through soil. Soil Sci. Soc. Am. Proc. 14:61-65. Reeves, R. B., D. W. Rennie, and J. R. Pappenheimer. 1957. Oxygen tension of urine and its significance. Federation Proc. 16:695-696. .Russell, M. B. .1952. Soil aeration and plant growth. Agronomy 2:254-501. Sawyer, D. T., R. S. George, and R. C. Rhodes. 1959. Polarography of gases: quantitative studies of oxygen and sulfur dioxide. Anal. Chem. 51:2-5. Sawyer, D. T., and L. V. Interrante. 1961. Electrochemistry of dissolved gases: II. reduction of oxygen at platinum, palladium, nickel and other metal electrodes. J. Electroanal. Chem. 2:510-527. Sawyer, D. T., and R. J. Day. 1965. Kinetics for oxygen reduction at platinum, palladium and silver electrodes. Electrochimica Acta 8:589-594. Taylor, G. S., and J. H. Abrahams. 1955. A diffusion equilibrium method for obtaining soil gases under field conditions. Soil Sci. Soc. Am. Proc. 17:201-206. Van Bavel, C. H. M. 1954. Simple diffusion well for measur- ing soil specific diffusion impedance and soil air composition. Soil Sci. Soc. Am. Proc. 18:229-254. 59 Van Doren, D. M., and A. E. Erickson. 1966. Factors affect- ing the platinum microelectrode method for measuring the rate of oxygen diffusion through the soil solution. Soil Sci. 102:25-28. Willey, C. R., and C. B. Tanner. 1965. Membrane-covered electrode for measurement of oxygen concentration in soils. Soil Sci. Soc. Am. Proc. 27:511-515. Wu, T. K. 1967. .Determination of dissblved oxygen concen- tration in soil with bare silver wire electrodes. M. S. Thesis. Michigan State University. Yamaguchi, M., F. D. Howard, D. L. Hughes, and W. J. Flocker. 1962. An improved technique for sampling and analysis of soil atmospheres. Soil Sci. Soc. Am. Proc. 26:512-515. ' APPENDICES APPENDIX A DESCRIPTION OF THE OPERATING CONTROLS AND TERMINALS OF THE OXYGEN CONCENTRATION METER This appendix gives a brief description of all the panel controls on the oxygen concentration meter. The controls are listed by their reference numbers as it appears in both Figures 8 and 9. 1. 2. 5. 4. 5. 6. 7. 8. 9. 10. 11. 12. 15. Positive Negative Positive External External External External ~External emf supply. emf supply. emf output to reference cell. power power power power power post for integrator, + 28.5 v. post for integrator, common. post for integrator, - 28.5 v. post for timers, + 25.5 v. post for timers, - 25.5 v. Power switch for emf supply. Power switch for the timers. "ON" uses internal batteries,. "OFF" connects to external batteries. Power switch for integrator. "ON" uses internal batteries, "OFF" connects to external batteries. Timer, set at 0.4 second. Timer, set at 0.1 second. 60 61 .Hmuma 00wumuuc00coo c00>xo mnu mo 20H00H0 Hmwuouowm .0.0Hsmflm NaomBUfiAm v . ® mmo .30 ~68ng m m 0 ago “no we momemoch seam mam . . use” UZOU MXO KOO QZ¢, ZMDmmmm exo on e 0 _ Hameeam nae new 0:8 mm BmHmmm had OMEN 9 0 0H a 0053 + mums: $d 00 we o e .. ....e mmo mmo mmo I +. 990 e e 8...... e e I :00 + 2H MQSHB MSW I hSH + ONE 0 mmm30m .BNm 62 .Houwa coflumnucoocoo c00>xo 000 no 80H00H0 mCHHflS ® 3 .0 wusmflm 0 Hum 00.00m Home 80 14. 15,16.‘ 17. 18. 19. 20. 21. 22. 65 Meter provides following full-scale readings. emf: 1.0 v Batteries: 40 v ODR: 20 wA Oxygen concentration, 0x1 : 1 v 0x2 : 0.5 v 0x4 : 0.25 v. Posts for measuring the resistance of the electrode systems. -emf output to preburn (5 minutes) an additional set of electrodes while the other set are in Opera- tion. The current can only be cut by disconnecting the plug at this position, and is not affected by the Operation of the timers and the switches 21, 24, 25 and 26. —emf output for oxygen concentration and O.D.R. measurements. emf adjustment. Battery selector--for checking voltages of 5 sets of batteries while (22) is at BAT, TIMER: check batteries for timers, 25.5 v. + INTEG: check batteries for integrator, + 28.5 v. ;_IN1§§: check batteries for integrator, - 28;5 v. Electrode selector. Selector-- BAT: For checking battery voltages 0x1, 0x2, 0x4: For oxygen concentration determina- tion. ~EMF: For checking emf. 25. 24. 25. 26. 27. ODR: For measuring oxygen diffusion rate. Zero adjustment for oxygen concentration measurement. Resistance compensator--This switch adds a resistance in the cell system. Preburn and ODR-~For 5 minutes preburn and ODR measurements. Keep at “OFF” in all the other cases. Oxygen concentration--Switch in "ON“ position it activates the timers. Discharge button--to discharge the feedback capacitor in the integrator and thus zeroes the meter. APPENDIX B OPERATING INSTRUCTIONS OF THE OXYGEN CONCENTRATION'METER The meter is designed for both oxygen concentration measurements and oxygen diffusion rate (O.D.R.) measure- ments. Collodion coated silver electrodes are used for the former and bare platinum electrodes are the latter. Front panel operation controls are shown in Figure 8, and the corresponding components are shown in Figure 9. IMPORTANT: To save batteries, switches g, 19, and 11 must be kept at "OFF" when the meter is not in use. I. The Measurement of Oxygen Concentration Check batteries 1. Turn.19_and 11 "ON". 2. Turn 2 to BAT. The full-scale reading is 40 v. 5. Turn 2 to ”TIMERS". The reading will be 25.5.v. Turn 2§_on. Two clicks will be heard from the timers, and the voltage drop should be less than 5 v. 4. Turn.2g to "+ INTEG". The reading should be no less than 26 v. The same should be satisfied when 29 is turned to "— INTEG". 5. Check the position of the balls of the wet cell. 64 65 Operation 1. 2. 5. 4. 5. 10. 11. 12. 15. 14. 15. 16. Turn §_and‘1Q_"ON", keep 11 at "OFF". Set 12 at 0.4 second and 1§_at 0.1 second. Set 21_at counter-clockwise limit. Set both 2§_and 26 at "OFF". Connecting electrodes to 18, Insert electrodes to the sample (1 cm above the place to be measured). Turn 21_to "PREBURN". Adjust emf to 0.65 v by turning 22_to emf and adjust 19, The full-scale reading is 1v. .Turn 22_to "0x2" (or "0x1"). Turn 2§_to "PREBURN", keep the preburn on for 5 minutes, then turn it "OFF". Insert electrodes 1 cm deeper, wait another 5 minutes to insure diffusive equilibrium. 'Repeat 6 while waiting. ‘ Turn 11_"ON", zero the meter by adjusting 25 while pushing 2.1.- Turn 21_to "1". Turn 2§_”ON". After hearing two clocks from the timers, turn 2§_"OFF". Record the meter reading. The reading can be doubled or halved by turn 22_to "0x4" or "0x11 respectively. Turn21_to the next electrode. Repeat 11 through 15 until all the electrodes are measured. .While electrodes are still in the sample, measure the resistances of the electrode system through pole ,15 and 16 with the Bouyoucos moisture meter. Place the same electrodes in air saturated 0.02 N.KCl. Force bubble the KCl solution well to insure air saturation. Repeat steps 6 and 7. Wait another 5 minutes. 17. .18. 19. 20. 21.. 22. 25. 66 Turn 21 to "1". Turn 17 to match the resistance in KCl solution to that in sample. Repeat steps 11 and 12. Turn 21 to the next electrode and repeat 18 through 20 until all electrodes are measured. Return 21_to counter clockwise limit. ~Divide the reading from 17 by the reading from 20 to calculate the standardized current. Oxygen concentration (in terms of oxygen satura- tion) can be obtained from the tOp curve in Figure 5 by interpolation. II. The Measurement of Oxvgen Diffusion Rate Turn 2,-19, 11, 2§_and 26 to OFF. Turn 21 to PREBURN and 2 to counter-clockwise limit. Insert the bare platinum microelectrodes into the soil and connect them to 12. .Turn on 2. Turn 22_to EMF and adjust the emf to 0.65 v (1 v full scale) with knob 12. Turn 22_to ODR. Turn 22 to ODR and wait 5 minutes. Take ODR readings of individual electrodes by turn- ing 21-