"OXIDATION-REDUCTION POTENTIAL (ORP) IN PREPARED AND
INDUSTRIALLY TREATED WATERS"

RAFAEL MARÍN GALVÍN (*), JOSÉ MIGUEL RODRÍGUEZ MELLADO,
MERCEDES RUIZ MONTOYA AND CARMEN JIMÉNEZ GAMERO

Departamento de Química Física y Termodinámica Aplicada
Facultad de Ciencias, Universidad de Córdoba
E-14004-Córdoba (Spain)
Phone: 34-57-218617/18/19; FAX:34-57-218606
E-mail:QF1ROMEJ@uco.es
(Received: March 2; 2000-Accepted: September 2; 2000)

ABSTRACT

A linear relationship between oxygen contents (as log %O₂) and E_H values for synthetically prepared waters has been obtained. Highly oxygenated waters showed values around E_H=-0.040V and rH=13.0, while poorly oxygenated waters showed values around E_H=-0.145V and rH=12.2. Such values were independent of the salinity. The presence of two active redox pairs typically encountered in natural waters such as "Fe³⁺/Fe²⁺" and "Mnⁿ⁺/Mn²⁺" also contributed to the E_H and rH values by increasing the ranges of variation with respect to those obtained in their absence.

In industrial drinking water production, dosage of O₃ or Cl₂ to raw water could serve to sterilise this water by increasing the E_H and rH values in 0.210V and 21.8 units, respectively.

In industrial biological treatment of wastewaters, very important reductions in the COD and NH₃ concentrations were associated to differences (0.300-0.400V in all cases from one water to another) in E_H in aerated water with respect to the E_H in the influent non-treated wastewater.

KEYWORDS: Drinking water. Redox potential. Synthetic waters. Wastewaters. Treated waters.

RESUMEN

Se ha obtenido una relación semilogarítmica entre la oxigenación del agua (como %O₂) y los valores de E_H para aguas preparadas sintéticamente. Las aguas con elevada oxigenación mostraron valores en torno a E_H=-0.04 V y rH=13.0, mientras que para aquellas poco oxigenadas los valores fueron de E_H=-0.145 y rH=12.2; estos valores fueron independientes de la salinidad. La presencia de dos sustancias redox típicas del agua como "Fe³⁺/Fe²⁺" y "Mnⁿ⁺/Mn²⁺" contribuyeron también a los valores de E_H y rH aumentando los rangos de variación respecto de los obtenidos en su ausencia.

En la producción industrial de agua potable, la dosificación de O₃ o Cl₂ al agua bruta puede servir para esterilizarla aumentando los valores de E_H y rH en 0.210 V y 21.8 unidades, respectivamente.

En el tratamiento biológico industrial de aguas residuales, se asocia la reducción muy importante en COD y concentración de NH₃ a diferencias entre los E_H del agua residual aireada y del agua residual no tratada. Estas diferencias fueron en todos los casos de 0.300 a 0.400 V en el valor de E_H de un agua a otra.

Palabras Clave: Agua potable. Potencial redox. Agua sintética. Aguas residuales. Agua tratada.

INTRODUCTION

Natural waters involve a great variety of redox pairs, being the H₂/H⁺ the most relevant for the ORP (oxidation-reduction potential). The redox potential of the reaction

½ H₂ = H⁺ + e^-

can be measured by using a normal H₂ electrode and an inert electrode (for example a platinum electrode). Taking into account the expressions of cell potential, the dissociation equilibrium of H₂ and the Nernst equation, it can be concluded that [1,2]:

E_H = 0.06log[H⁺] - 0.30logP_H2

and considering with Clark [3] that "log(P_H2)=-rH", it can be established that [1, 2, 4]:

rH = E_H/0.30+2pH

This equation relates the two parameters commonly used to express the ORP in water, "rH" and "E_H". High values of rH (>15) and E_H of positive sign (or very low E_H values irrespective its sign) characterize oxic conditions, and conversely, low values of rH and E_H of clearly negative sign are characteristic of anoxic waters [1, 2, 4].

The specific redox state of a given water (natural, wastewater or along its treatment processes) drives its behaviour and the concentrations of different compounds [4-14]. Two circumstances can be found for a water: high oxygenation (high ORP values) and abundance of oxidised species (i.e. Fe³⁺ and Mn⁴⁺ slightly soluble, inorganic carbonaceous compounds, sulphates, nitrates...) and deficient oxygenation (low ORP values) and abundance of reduced species (i.e., Fe²⁺ and Mn²⁺ sparingly soluble, organic carbonaceous compounds, methane, sulphides, nitrites...). Such situations succeed one to another depending on the stational changes of the environment for natural waters, or the stage of manipulation for treated, drinking and waste water. In addition, the lower or higher ORP strongly influence the subsequent industrial treatment processes of these waters.

The aim of this paper was to contribute to the study of the ORP in waters in two ways: first, the relation between oxygenation, ORP and concentrations of Fe and Mn for prepared synthetic waters has been studied. Second, two practical applications of the ORP evolution of water along the treatment processes have been made, with the aim to optimize the industrial treatment of water: (a) raw natural water intended for water drinking production and (b) wastewater intended for biological treatment.

MATERIALS AND METHODS

Measurements of the potential vs. time curves were carried out by using a METEX M-6450 CR digital multimeter coupled with a MULTITECH ACER 710 computer. The working electrode was a Metrohm 6.0302.100 Pt electrode and a saturated calomel electrode Metrohm AG 9100 was used as reference electrode. All potentials were referred to the normal hydrogen electrode by subtraction of the potential of the saturated calomel electrode. The electrolytic cell used was thermostated and equipped with a pipe to bubble nitrogen or air into water. Oxygen was measured with a CRISON OXI92 oxymeter with temperature compensation and equipped with a CRISON E090 oxygen electrode.

The analysis of waters was made according to the "Standard Methods" [12]. Measurements of UV absorbance were recorded by using a computerised UV-Visible Perkin-Elmer Lambda 3b Spectrophotometer, after samples centrifugation at 5000 rpm during 5 min, to clarify the liquid [4]. Finally, the total number of coliforms present in the waters was obtained by the usual membrane filter procedure [12]. Synthetic waters were prepared according to the "theoretical composition" method as described elsewhere [1]. Thus, three kinds of synthetic waters were obtained: low salinity, middle salinity and high salinity waters (see table-1). Concentrations of Fe²⁺, Fe³⁺ and Mn²⁺ were obtained by dissolving FeCl₂∙4H₂O, FeCl₃∙6H₂O and MnCl₂∙4H₂O, respectively.

Table I. Preparation of synthetic waters*

Waters under treatment to drinking water production were taken during October-November 1994 in the Villa Azul Drinking Water Plant which supplies to Córdoba City (Spain). This plant produces 150.000 m³ drinking water/day (maximum), the treatment consisting during the study in aeration of the raw water, ozonization, prechlorination with Cl₂, settling, filtration through rapid sand filters and postchlorination (Cl₂) [13, 14]. Forty-eight samples were taken, distributed in the groups "raw water", "oxidized water" and "filtered water". Water samples without treatment influent to plant were named "raw waters"; "oxidized waters" corresponding to raw waters aerated, ozonized (1-1.5 g/m³ or ozone) and prechlorinated (4.5 to 5.2 g/m³); "filtered waters" were the above ones settled with alum (dosed between 28 to 30 g/m³) and filtered through siliceous sand rapid filters.

Wastewaters under biological treatment were taken during September-October-November 1994 in La Golondrina Wastewater Treatment Plant which treat the total (domestic and industrial) waste waters of Córdoba City. The plant must treat up to 108.000 m³/day and the treatment employed was the following: double screening, removal of sands and greases-oils, primary settling, aeration, secondary settling and sludge treatment (incineration or chemical treatment with Ca(OH)₂) [15]. Fifty-seven samples were studied distributed in the series "waste water", "settled water" and "treated water". Waste water samples without treatment influent to plant were named "waste waters"; "settled waters" corresponded to screened waste waters, without sands and greases and primarily settled; "treated waters" finally, were the above ones aerated (to a residual oxygen concentration of 1.5 g/m³) and finally settled after biological treatment.

Two methods were used to obtain waters with a given oxygen concentration:

(a) The sample is saturated with oxygen by bubbling O₂ (or, alternatively, air) during 8-10 min. The oxygen is then slowly removed from the water by bubbling N₂.

(b) All the O₂ is removed from the water by bubbling N₂ and the oxygen concentration is slowly increased by bubbling O₂ or air.

In both cases the oxygen content is monitored with the oxymeter and when the desired concentration is reached, the water is used.

RESULTS AND DISCUSSION

The rate of electron transfer towards a platinum electrode is associated to the surface characteristics of the Pt. After each measurement adsorbed substances cover the electrode preventing new measurements [12]. For this reason, the electrode exhibits a "memory effect" which strongly affects the measurements of the following samples. Several cleaning methods were used to overcome these problems [1, 9, 16-18]. Among these methods, and after a variety of essays, we have selected the following cleaning procedure before each measurement: introduction of the Pt electrode during one minute into sulfo-chromic mixture (6 g of K₂Cr₂O₇+100 ml of concentrated H₂SO₄ in 250 ml of total dissolution) and rinsing with abundant distilled water. So, the potential vs time curves and the E_H and rH values obtained were highly reproducible.

(a)Synthetically prepared waters

To study the influence of oxygen on waters ORP, different oxygen concentrations were taken for prepared waters according to method (a) described in the precedent section. Once the electrode is immersed into the water, the recording of ORP starts. This recording was made in a continuous mode, maintaining the electrode into the water until the end of the experiment. The ORP varied with the measurement time in the general shape shown in Figure 1, that is, reaching a limiting value at long times. The establishment of the end value of ORP is driven by the diffusion of charged species towards or from the electrode and by the adsorption/desorption of such species on the electrode surface, explaining (at least qualitatively) the shape of the E_H vs. time curves. This general behaviour was found for all synthetically prepared waters, irrespective their composition. Thus, E_H is the "equilibrium oxidation-reduction potential" and corresponded to ORP when the equilibrium in the interface is reached. It must be remarked that the measurement corresponded in all cases to a unique oxygen content and the stabilisation time was reached when the E_H values changed less than 5 mV (being this quantity the accuracy in E_H), in this case after 1300 seconds. Table-2 presents the data of %O₂, E_H, rH and pH for all the prepared waters investigated. As can be seen, higher oxygenations implied higher E_H and rH values. Conversely as occurs for a natural aquatic medium, the lower oxygenation implied the higher pH-values. This fact can be explained by taking into account the method used to obtain the actual oxygen concentration: the O₂ removing by bubbling N₂ (or vice-versa) implies a parallel decreased of CO₂ in the medium (this gas is also removed) and, consequently, the increase of pH.

Fig. 1. Potential vs time curves from middle salinity prepared waters. %O2 saturation from down to up: 8, 19, 31, 47, 61, 75 and 86.

Tabla II. Results obtained for synthetics waters*

In prepared waters only the O₂ influences significantly both E_H and rH values as can be inferred from the linear correlation between E_H and log %O₂ shown in fig.2. Moreover, synthetic waters with >70% in O₂ saturation showed rH values around 13 in accord with the experimental values checked for unpolluted natural waters [1, 2, 9, 19].

Fig. 2. Prepared waters: relation EH vs log %O2 saturation. White circles: high salinity waters; black circles: middle salinity waters; triangles: low salinity waters.

The potential vs. time curves were also studied for prepared waters after addition of Fe²⁺, Fe³⁺ and Mn²⁺ ions at the following concentrations (in mg/l): 0.1, 0.3, 0.6 and 1.0. Waters with Fe³⁺ were firstly saturated in O₂ and then slowly deoxygenated by bubbling N₂ (method a), while waters with Fe²⁺ or Mn²⁺ were firstly saturated in N₂ and then oxygenated by bubbling O₂ or air (method b).

Addition of Fe²⁺, Fe³⁺ and Mn²⁺ caused variations in the potential vs time curves with respect to the shapes of the curves obtained in the absence of these ions. The magnitude of the distortions followed the sequence "Mn²⁺>Fe²⁺>Fe³⁺". Likewise, the higher the concentration of the ions, the higher the distortion. This can be explained by taking into account the high adsorption of oxihydroxides of Mn and Fe generated during the experiments that harmed the diffusion processes associated to the kinetics of the measurements of ORP. Moreover, the number of possible oxihydroxides generated for Mn is higher than those expected for Fe [20-22], explaining the above sequence.

On the other hand, the more oxygenated samples showed always the highest rH values and the more positive (or less negative) E_H values, with independence of salinity and concentrations of ions (table-2). This could be associated to the reversibility of the process as occurs in nature.

The range of variation of E_H values in the samples containing Fe or Mn ranged more widely than in the rest of the samples. This must be related to the occurrence of the new redox pairs "Fe³⁺/Fe²⁺" or "Mnⁿ⁺/Mn²⁺" in addition to the redox pair oxygen-water, which undoubtedly contributed to the waters ORP value.

(b)Waters under treatment to drinking water production

Potential vs. time curves corresponding to these waters were different from those corresponding to prepared waters in the absence of Fe or Mn. This can be explained by the presence of organic matter that show a strong tendency to adsorption on the Pt electrode surface. In this case the equilibration time for ORP, O₂ and pH measurements was 1800 seconds.

Table-3 shows the evolution of the water characteristics along the treatment for drinking water production. Important reductions were observed in Fe, Mn, coliforms, UV-254 nm absorbance and COD-Mn (organic matter to KMnO₄) from raw untreated water to final drinking water.

Table III. Characteristics of water under treatment for drinking water production (maximum and minimum values)

The evolution of E_H and rH in waters under treatment follows the sequence: "oxidized">"filtered">"raw waters" (see fig.3). The oxidized water presented the higher values due to the dosage of some oxidants (O₃ and Cl₂) specially in the oxidation-prechlorination step. The residual oxidant was lost by both oxidation and atmospheric diffusion, implying the decrease of the above values in filtered waters with respect to oxidized waters. The raw waters showed the lowest E_H and rH values of each group due to the oxidant power of O₃ and Cl₂, much higher than the air oxidant power. However, the raw waters showed always ORP values corresponding to oxic waters [1, 9, 19].

Fig. 3. Waters under treatment for drinking water production: variation of ORP along the treatment (average values). (A) EH; (B) rH.

Oxidation processes used in water treatment originate the removal of oxidizable species (dissolved iron and manganese, organic matter, etc..). The efficiency of this process must be associated to both the specific dose of oxidant employed and the increase of water ORP. In this study, the dosages used for both O₃ and Cl₂ were relatively low. Moreover, the raw waters untreated showed a reasonably good oxidation state and low levels of Fe, Mn and organic matters. For this reason, it was difficult to prove an experimental relation between high removing of substances easily oxidized, present in water, and high increases of E_H and rH in oxidized waters against raw waters. According to these results, the ORP value is not an adequate parameter for controlling the treatment process in production of water intended for human consumption: it is more reasonable to work by optimizing the oxidant dosages as it is commonly made.

Disinfection of water is made with the same reactants as oxidation, the oxidant playing two roles: chemical action with damage of cellular wall and interference on specific enzymatic reactions for the metabolism of microorganisms [23]. Previous papers [23, 24] showed that the increase of E_H in waters associated to the dosification of Cl₂ and/or O₃ was related to the effectiveness in removing escherichia coli and other microorganisms. Increases of 0.5-0.6V were sufficient to promote the disinfection in less than 30 minutes. In the present study it has been observed the complete elimination of the total coliforms number present in raw waters for increases of E_H and rH values in oxidized water against raw water equal to 0.21V and 21.8 units, respectively, lower than those reported in the literature. This can be due to the low microbiological amount in our waters against the higher used in other studies. Moreover, such previous studies considered only laboratory experiments and not practical industrial cases as that studied here.

(c)Waste waters under biological treatment

As occurred for the above mentioned cases, potential vs. time curves presented distortions which can be probably due to the very important amounts of substances present in waste waters, specially organics, which are strongly adsorbed on the platinum electrode surface. Nevertheless, approximately after 1800 seconds the equilibrium was established.

Table-4 shows the evolution of waste water characteristics along the biological treatment: high amounts of COD, BOD₅ and NH₃ were removed from untreated waste water. Particularly, the E_H and rH values followed the sequence: "treated water">"waste water">"settled water" (see fig.4). This can be explained by taking into account that the treated water experienced a process of aeration prior to the biological treatment being relatively well oxygenated and thus relatively oxic. Settled water did not experienced aeration, its very low O₂ level was consumed and, in consequence, its reduction state was increased with respect to the wastewater.

Finally, waste water was originally anoxic as is typical in these kind of waters. Nevertheless, the E_H and rH values measured in this study were lower than those found in the literature [1, 2, 4, 9] for recently produced waste waters. This can be due probably to the consumption of O₂ in these waters along the sewage network pipes that carry waste water to treatment plant.

Fig. 4. Waste waters under biological treatment: variation of ORP along the biological treatment (average values). (A) EH; (B) rH.

Fig.5 shows a non linear relationship E_H vs. %O₂ for these wastewaters. For waters very poorly oxygenated, (%O₂<10%) a wide variation in E_H values (-500 mV to -100 mV) can be observed. This has not an evident explanation but it can be assumed that under these conditions other redox species different from O₂ can influence the redox state. Among such species, the equilibrium "sulphate-sulphide" could be very outstanding in this case, as well as all the reduction processes used by the waste waters anaerobic microorganisms [4, 8, 25, 26]. For medium-high O₂ values (%O₂>40%), a linear relationship between E_H and %O₂ was achieved, indicating that anaerobic processes do not influence the redox state of water.

Fig. 5. Waste waters under biological treatment: relation between EH and %O2 saturation for all the available data.

Domestic waste waters contain high amounts of residual organic substances and are poor in dissolved O₂. Consequently, a relationship between organic load and both O₂ and reductive state in wastewaters can be expected. Fig.6 confirms that waters with high BOD₅ values showed low rH and negative E_H values.

Fig. 6. Waste waters under biological treatment: relations BOD5 vs rH.

Biological treatment of domestic waste waters is based on the controlled growth of an aerobic microorganisms culture. This is carried out through water aeration before the secondary settling, in our case up to a residual O₂ concentration of 1.5 g/m³. In this way, the highest percentages in organic substances removed from water must be associated to an optimal oxygenation process, and then to a sufficiently low negative E_H value or to a sufficient increase in the rH value of aerated water against not aerated water. Thus, decreases in COD and NH₃ of water around respectively 450 mg/l and 8-10 mg/l (in absolute values) were statistically associated to increases of 0.300-0.400 mV of the E_H value in aerated water against influent waste water.

These variations can be obtained by modifying the air dosage to water, which should promote modifications of the water ORP, better than by operating the plant to obtain a pre-fixed residual O₂ concentration. This is used in many industrial plants of biological treatment, especially in anaerobic treatment and to reduce nutrients in waste waters, with high efficiency [27, 28] and could imply important savings in the exploitation of the plant.

CONCLUSIONS

(a) Synthetic waters

- A linear relationship between the logarithm of oxygenation and ORP was obtained, with independence of water salinity: E_H ranged between -0.040V and -0.145V and rH between 13.0 and 12.2 units for very oxygenated and poorly oxygenated waters, respectively.

- Addition of Fe³⁺, Fe²⁺ or Mn²⁺ to prepared waters implied variations of the ORP due to the contribution of the redox pairs Fe³⁺/Fe²⁺ or Mnⁿ⁺/Mn²⁺. Thus, E_H and rH ranged, respectively, from -0.020V to -0.150V and from 15.0 to 12.0 units for very oxygenated and poorly oxygenated waters, respectively.

(b) Raw waters and along treatment to drinking water production

- Relations between ORP and oxygenation have been stated for both untreated raw waters and waters under treatment for drinking water production. In all cases, the higher the oxygenation, the higher the ORP measured.

- For waters under treatment for drinking water production the ORP was strongly influenced by the oxidant used, O₃ or Cl₂. Since such oxidants are much more oxidant than air, the measured ORP increased.

- Disinfection of raw waters was observed with associated increases in the E_H and rH values lower than those reported in the literature. Increases of 0.210V (E_H) and 21.8 units (rH) in oxidized waters compared to raw waters guaranteed the complete elimination of total coliforms. Nevertheless, the practice indicates that industrial disinfection of water must be carried out always by optimising the oxidant dosages and not operating with any pre-fixed ORP value.

(c) Untreated and along biological industrial treatment waste waters

- A non-linear relationship between E_H and percentage of dissolved oxygen has been found for both untreated and under biological industrial treatment waste waters. Low oxygen contents (less than 10-15%) implied a wide variation in E_H values, whereas waters with 40% O₂ or more presented a small E_H change with oxygenation.

- A correspondence between organic load (expressed as BOD₅) and ORP has been proved. The more negative E_H values or the lower rH values were statistically associated to the higher BOD₅ values in the range of 13-470 mg/l of BOD₅.

- Finally, important reductions of COD (-85%) and NH₃ (-28%) concentrations were associated to increases of 0.300-0.400V in the E_H value of aerated water against influent wastewater. This can indicate that the biological industrial process could be optimized better by adjusting the ORP values measured as a function of the yield instead of by the conventional method to obtain residual pre-fixed O₂ levels.

ACKNOWLEDGEMENTS.

Authors wish to express their gratitude to the Empresa Municipal de Aguas de Córdoba S.A. for supplying the industrial water samples; to Junta de Andalucía for partial financial support and to the Spanish Ministerio de Educación y Ciencia for a postdoctoral Grant (Grant IN92-B30528630).

REFERENCES

1. Rodier, J. (1989) In: Análisis de Aguas. Omega S.A., Barcelona.

2. Catalán Lafuente, J. (1981) In: Química del Agua. Alonso, Madrid.

3. Clark, W.M. (1961) In: Oxidation-reduction Potential of organic systems. Williams and Wilkins, Baltimore.

4. Marín Galvín, R. (1995) In: Análisis de Aguas y ensayos de tratamiento: Principios y Aplicaciones. G.P.E. S.A., Barcelona

5. Odum, E.P. (1971) In: Fundamentals of Ecology, 3th. ed. Saunders Co., New York

6. Stumm, W. and Morgan, J.J. (1980) In: Aquatic chemistry. Wiley Interscience, New York

7. Wetzel, R.G. (1981) In: Limnología, spanish ed. Omega S.A., Barcelona

8. Tchobonablous, G. and Schroeder, E.D. (1985) In: Water Quality, vol. I. Adisson-Wesley Pub. Co., Reading (Mass.)

9. Morel, F.M.M. and Hering, J.G. (1993) In: Principles and Applications of Aquatic Chemistry. Wiley and Sons Inc., New York

10. Marín Galvín, R., European Water Pollution Control, 3(3), 15-20, (1993)

11. Marín Galvín, R., European Water Pollution Control, 3(4), 36-43, (1993)

12. AWWA. Standard Methods for the examination of Water and Wastewater, 18th edn. American Public Health Assoc., Washington D.C. (1992).

13. Marín Galvín, Water Research, 26(5), 663-668 (1992)

14. Marín Galvín, R. and Rodríguez Mellado, J.M., Water S.A., 19(3), 231-234 (1993)

15. Empresa Municipal de Aguas de Córdoba S.A. (EMACSA) Annual Reports 1989-1993 and other technical informations.

16. Koryta, J. and Faulkner, L.R. (1987) In: Principles of Electrochemistry. Wiley and Sons Inc., New York

17. Yu, T.R. and Ji, G.L. (1993) In: Electrochemical methods in soil and water Research. Pergamon Press, Oxford

18. Genders, D. and Weinberg, N. (1993) In: Electrochemistry for a cleaner environment. Electrosynthesis Co. Inc., New York

19. J.M. Rodríguez Mellado and R. Marín Galvín (1999) Fisicoquímica de aguas. Ed. Díaz de Santos. Madrid. Spain

20. Pourbaix, M. (1966) In: Atlas of electrochemical equilibria in aqueous solutions. Pergamon Press, Bruxelles

21. Grundl, T. and Malacady, D.M., J. Contam. Hydrol., 5, 97-117 (1989)

22. Grundl, T. and Delwiche, J., J. Contam. Hydrol., 14, 71-79 (1994)

23. Carlson, S., J. Water SRT-Aqua, 40(6), 346-356 (1991)

24. Clark, W.M. and Cohen, B. Public Health Reports, 38, 666 (1923)

25. Nemerov, N.L. (1977) In: Aguas residuales industriales, teoría, aplicaciones y tratamiento. Ed. Blume, Barcelona

26. Schroeder, E.D. In: Water and Wastewater treatment. McGraw- Hill, New York, 1977

27. Guardino Ferré, R. and Zorrilla Soriano, F. Obras Públicas, 31, 80-83 (1995)

28. Hernández Muñoz, A. (1992) In: Depuración de Aguas Residuales, 2nd edn. Chap. 11th, pp. 791 Digestión Anaerobia. Ed. Esc. Sup. Ing. C.C.P., Madrid

---

(*) To whom all correspondence should be addressed.