Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Langmuir. Author manuscript; available in PMC 2010 March 26.
Published in final edited form as:
PMCID: PMC2845823

Can Water Store Charge?


Previous work from this and other laboratories has demonstrated large pH gradients in water. Established by passing current between immersed electrodes, pH gradients between electrodes were found to disappear slowly, persisting for tens of minutes after the current had been turned off. We find here that these pH gradients reflect a genuine separation of charge: at times well after disconnection of the power supply, current could be drawn through a resistor placed between the charging electrodes or between pairs of electrodes positioned on either side of the midline between original electrodes. In some experiments, it was possible to recover the majority of charge that had been imparted to the water. It appears, then, that water has the capacity to store and release substantial amounts of charge.


Electrolytic processes in pure water are commonly ascribed to the following sequence of events: current flow, electrode polarization, interfacial double-layer formation around the electrodes, and establishment of equilibrium.1 The charge carriers responsible for this sequence include negatively charged OH groups, protons, hydronium ions, and perhaps larger scale charged clusters.2

In a previous report from this laboratory,3 it was found that the distribution of charge carriers did not fit this conventional picture. During the passage of current through a chamber filled with water and pH-sensitive dye, large pH gradients were created. The region next to the anode developed extremely low pH, while the region next to the cathode developed extremely high pH. Each of these zones of approximately uniform color spread throughout roughly one-half the chamber, leaving a steep gradient over a narrow strip near the center.

In this report, we amplify and extend these observations. We find that the applied currents cause appreciable amounts of charge to be stored in the water, charge that can later be extracted in the form of current flow. It appears that the previously reported zones of high and low pH correspond, respectively, to regions of net negative and positive charge.


The experimental setup consisted of an electrochemical cell with electrodes positioned at either end, along with simple circuitry that supplied voltage and allowed the readout of certain electrical variables (Figure 1).

Figure 1
Schematic model of experimental setup.

The experimental protocol involved two phases, charging and discharging. During charging, the switch in Figure 1 was connected to point 1, so the supplied current passed through the electrodes and into the solution. The DC power-supply voltage was set somewhere between 2.5 and 4 V. For discharging, the switch was connected to point 2. From this moment on, the electrochemical cell itself worked as a power supply, discharging any stored energy into a resistive load. In some protocols, as noted, an extra pair of discharge electrodes was added to, or substituted for, the original electrode pair.

The electrochemical cell was fabricated from a standard disposable plastic spectrophotometer cuvette, with inner dimensions 5.4 × 1.0 × 1.0 cm. The plastic material is transparent within the spectral range up to 800 nm. The cuvette’s open end was closed by sheet of plastic cut from a similar cuvette and held in place by silicon glue. Two slots for plate electrodes were cut parallel to the chamber’s opposite ends (Figure 1). For experiments where alternative pairs of electrodes were used for discharging, a series of hole-pairs were drilled into the top face of the cell along the midline and equidistant from the slots (Figure 1). For early experiments, a similar cuvette, but with dimensions 2.6 × 1.0 × 1.0 cm, was used and contained slots to accommodate only one pair of electrodes.

The primary electrode pair, which was used for charging, and in some experiments also for discharging, was made from a 1.4 cm × 0.9 cm piece of 0.127 mm platinum foil. A second pair of electrodes, used in the charge-distribution-defining experiments, consisted of platinum wires 0.25 mm in diameter. Metal for all electrodes was specified at 99.99% purity.

Because pure water is a poor conductor, in most experiments a dilute salt solution (NaCl < 10 mM) was used. However, control experiments were carried out in the absence of salt, with ultrapure water (Type 1 by ASTM’s [American Society for Testing and Materials] D1193-06 standard) or distilled water. Also, pH-sensitive dye was employed for distinguishing regional pH values. For this purpose, a universal indicator (Riedel-de Haen 36803 pH 3.0–10.0) was used in concentrations recommended by the manufacturer, three drops of liquid dye per 10 mL of solution, or approximately 1.5% v/v.

In early experiments, electrical parameters were recorded using a simple digital multimeter (TENMA, part # 72-7745 RC). When currents were too small to resolve, especially during discharging, they were calculated from the value of Rload and the observed voltage readings across that resistor.

In later experiments, all results were obtained using a data-acquisition device (DAQ), manufactured by National Instruments (NI USB-6009). The DAQ served both as a signal source and as recorder of voltage and current. With software written in Labview, the values of charging voltage, period of charging, period of discharge recording, sampling rate, and data-file name could all be set. To compensate for the relatively low input impedance of the DAQ-input lines, high input impedance buffers were added between the electrodes and the DAQ. Data were automatically uploaded into a computer when the experiment was completed. For Rload, a fixed resistor was used.

For control experiments on pH distribution during and after electrolysis, the setup depicted in Figure 2 was used. It consisted of two flasks, each containing 2.5 mL of water, with an interconnecting water bridge made of silicon tubing with 3 mm inner diameter. To pass current, a platinum electrode was immersed into each flask.

Figure 2
Setup for control experiments on pH distribution during and after electrolysis.


Charging Dynamics

Dynamics of charging were studied using the setup of Figure 1. The electrolyte was either distilled water or 10 mM NaCl, as indicated. The DC power supply was set to 4 V. Representative plots of current versus time are shown in Figure 3.

Figure 3
Dynamics of charging phase: Potential difference of 4 V was applied to plate electrodes immersed in 10 mM NaCl electrolyte solution (top) or pure water (bottom).

The results shown in Figure 3 were found consistently in 12 such experiments. Typical curves of current versus time showed three phases: (i) a rapid initial falloff; (ii) recovery, with continued growth; and (iii) steady current flow, continuing until the power supply was turned off. When pure water was used instead of the NaCl electrolyte, results were qualitatively similar; however, current magnitude was lower, and transitions between phases were less obvious.

Several experiments were carried out with an electrochemical cell shorter in length, 2.6 × 1.0 × 1.0 cm. With this shorter cell, both volume and distance between electrodes were reduced. Out of six experiments with 10 mM NaCl, the current plot was displaced downward by approximately 0.07 mA as compared to the standard curve plotted in Figure 3. Otherwise, results were similar.

The plot of current versus time differed in nonlinear fashion when the input voltage was altered beyond the standard window. When the applied voltage was reduced to less than 1.5–2 V, the initial phase (i) showed low current, which declined to zero and never recovered to any detectable level. When the voltage was increased above the standard range of 2.5–4 V, the current plot shifted upward, while both the minimum point and the onset of phase iii shifted rightward.

The dynamics of pH change during charging, more thoroughly described in a previous report,3 are shown in Figure 4. Significant changes of pH uniformity occur during the first several minutes after initiation of current flow. Within 10 min, clear differentiation is established between zones of high and low pH. Beyond that time, changes are relatively minor.

Figure 4
Water–pH dynamics during charging, recorded from above the chamber. Electrodes situated along the top and bottom of each panel. Times after current-flow initiation are indicated below. The colors of the indicator correspond to the following values ...

Control experiments were devised to test whether the pH separation might arise as a result of charging the dye molecules instead of the water. In these experiments, current was passed in the absence of dye, using the control setup of Figure 2. This arrangement allows physical separation of zones, so that when the dye is subsequently added, there is minimal disturbance and the distinct zones cannot easily mix.

To carry out this experiment, flasks and bridge tubing were filled with water. A potential difference of 4 V was applied to the electrodes for one hour. The circuit was then opened, and tubing and electrodes were immediately removed from the flasks. Next, 50 µL of universal indicator was added to each vessel. The color in one vessel corresponded to pH 8 and in the other vessel to pH 4. This result was found consistently (n = 4). Thus, large pH differences are noted even without dye present during charging. (For reference, a pH value of 4 corresponds to 10−6 protons per molecule of water.)

In Figure 4, a bubble can be seen growing in the cell’s upper left corner. Bubble formation occurred in all experiments (n > 20), although position and growth rate were inconsistent. In most cases, formation began during the charging phase and continued through discharge. Characteristics of bubble formation were not pursued in any detail, but may warrant future study.

Discharge Dynamics and Stored Charge Distribution

The dynamics of pH distribution were visualized also during discharge. For these experiments, the charging phase was set to 45 min. Figure 5 shows representative examples of the pH distribution during the period of discharge in two situations: when the circuit between electrodes was open (top) or closed (bottom). In both cases (n = 3), zones of different pH remained for extended periods, on the order of 1 h. Moreover, the “red” (positive) zone barely changed in terms of depth of color until the other zone had become almost neutral.

Figure 5
pH dynamics during discharge, which began at time 0. Top: Open circuit between electrodes. Bottom: Short circuit between electrodes.

Because pH zones persist for long times, implying that the attendant charges may be stored throughout the chamber, the question arises: Can at least part of the charge be recovered? To address this question, additional sets of electrodes were used to first map the charge distribution. The setup and charging processes were similar to those described above. During discharge, data were collected with additional electrode pairs prepositioned at each of a series of locations throughout the cell. These positions were selectable along the midline of the chamber’s long side. Hence, the distance between electrodes, as well as distance from the original electrodes, could be regulated. During the charging phase, the circuit between these additional electrodes was left open, and, during discharge through these electrode pairs, the electrode pair used for charging was also left open circuited.

The results are shown in Figure 6. The voltage obtained from electrodes positioned closer to charging electrodes was greater than that obtained closer to the chamber’s center. This feature applied over the short-term (Figure 6a) as well as the long-term (Figure 6b). Because electrode pairs were positioned at fixed points fairly widely separated, the voltage distribution could not be determined with high spatial resolution. Nevertheless, the correlation between discharge voltage and position was consistent in 15 experiments. Although the apparent charge stored near the center was less than that stored near the ends (near the charging electrodes), it is notable that some charge was stored throughout the full extent of the chamber.

Figure 6
Time course of discharge voltage at various electrode positions. Second member of electrode pair placed symmetrically on right side. 4 V charging voltage; 1 h charging time; 10 kΩ load resistance. 10 mM NaCl electrolyte. (a) Initial 10 s from ...

To ensure that the results were not an artifact stemming from residual charge stored or collected on the charging electrodes, control experiments were carried out. The strategy was to remove charging electrodes after charging and place the discharge electrodes in the chamber only after their removal, 10 min after charging was complete.

Figure 7 shows the results of three separate experiments along these lines. Plot (1) is the reference curve, with both charging and discharging electrode pairs in place throughout. Plots (2) and (3) show data obtained from additional electrodes added 10 min after charging had ended, the latter obtained with charging electrodes removed as the discharging electrodes were added. The results demonstrate substantial potential difference between added electrodes even after the charging electrodes had been removed. The gap between curve (1) and the others may have arisen from the disturbance attendant with electrode additions/replacements or from the charge that may have accumulated in discharge electrodes during the charging process. Or, because electrode area may play a role, it is possible that the discharge voltage might have been greater if plates had been used instead of wires, for the additional electrodes. At any rate, substantial charge is evidently stored throughout the chamber.

Figure 7
Control experiments. 4 V charge voltage, 1 h charge time, 10 kΩ load resistance. 10 mM NaCl electrolyte. Discharge electrodes located 9 mm from the charging electrodes (third hole).

Charge Recovery

From experiments such as those illustrated in Figure 6, it is possible to calculate the ratio of recovered charge to input charge or energy-recovery efficiency. Because the discharge current in the protocols above was considerably smaller than the charge current, we decided to explore efficiency on a shorter charge-time scale.

Figure 8 shows representative results obtained with a single pair of plate electrodes. The system had been charged for just 3 s; however, the potential difference between electrodes persisted for more than 2 min following power-supply disconnection. From such records, one can integrate the current over time to obtain the charge recovered. Similarly, integration of input current yields input charge, and the ratio of the former to the latter yields efficiency.

Figure 8
Discharge-voltage dynamics measured after disconnection from DC power supply. Charging voltage 2.5 V, applied for 3 s; load resistance, 10 kΩ.

To explore the domain of efficiency, the field of parameters that varied is evidently vast. Over the relatively limited set of conditions explored, with plate-charging electrodes, we found that the maximum efficiency occurred with the following parameter values: charging time, 1 s; charging voltage, 2.5 V; load resistance, 0.75 kΩ; solution composition, 10 mM NaCl. Under those conditions the recovered charge was 0.14 ± 0.01 mC, while the input charge was 0.20 ± 0.01 mC (n = 5). Consequently, the recovery ratio was approximately 70%. This is not necessarily the absolute maximum, but more likely a local maximum realized over the limited domain of conditions explored.

The ratio did not depend strongly on whether the solution was pure water or water with modest salinity. A representative result is shown in Table 1, obtained under conditions only slightly different from those described above: single pair of wire electrodes used for both charging and discharging; charging voltage 2.5 V; charging time 3 s; load resistance 10 kΩ. The table shows that, although the absolute values of charge increased by approximately three times in the presence of salt, the ratio remained approximately the same.

Table 1
Recovery Ratio (n = 5) with and without Salt

Use of plate electrodes instead of wire electrodes improved the ratio considerably. Under standard conditions (2.5 V charging voltage for 3 s, 10 kΩload resistance, 10 mM NaCl), the amount of charge delivered to the cell was about the same with wire or plate electrodes, 0.26 ± 0.01 mC (n = 5), but the charge that could be extracted was approximately three times higher with the plates than with similarly positioned wires, 0.17 ± 0.01 mC versus 0.06 ± 0.01 mC.

The presence of pH-sensitive dye also made some difference. When the salt solution contained pH-dependent dye at the concentration recommended by manufacturer, input and output charges were, respectively, 0.24 ± 0.01 and 0.08 ± 0.01 mC (n = 5), a ratio of about 0.32. The same conditions, but without the dye, gave 0.23. Hence, the presence of dye had a modest effect on the results.


The present experiments were carried out to explore the possibility of charge storage in, and subsequent recovery from, water. Previous studies demonstrated long-term charge separation following electrolysis,1 and this possibility had been noted in other studies.4,5 Charge storage is also dramatically implied by the classic Kelvin Water-Dropper experiment, an excellent manifestation of which is presented at Droplets accumulating in proximate metallic chambers build potential difference large enough to create visible discharge between chambers, implying appreciable charge storage in each body of water.

What is especially surprising in the current results is that the separation of charge within the chamber could be maintained for long periods of time. Extensive zones of distinct pH remained in place for up to several hours following cessation of charging (Figure 5). Because these distinct zones correspond to zones of positive and negative charge (Figure 6), the expectation is that they should readily mix, but they do not neutralize easily. The situation is much like a battery in which charges can remain separated for long periods and can then be drawn on demand.

Electrode-surface effects are well recognized.6 However, the separated charges appear to be stored in the water itself and not on the electrode surfaces. This was determined by removing the charging electrodes and replacing them with fresh electrodes: the potential difference between half-chambers remained evident (Figure 7). Although the potential difference was lower closer to the chamber center, an appreciable potential difference nevertheless remained even very near to the center (Figure 6). Hence, charge is stored over each entire half-chamber and must therefore lie within the water itself. It is apparently not an electrode-based phenomenon.

Perhaps the most significant aspect of the results is that stored charge could be recovered. This was shown by placing a resistive load between electrodes and drawing current through this load. The recovered-current magnitude depended on many factors. Although no systematic attempt was made to map the range of recovery over the full domain of possible factors, we did find that recovery as high as 70% of input charge was not difficult to obtain. This implies that, as a storage medium, water is quite effective.

A number of fundamental questions are raised by these results. One of them is why separated charges in solution can remain separated for periods as long as those observed. Separated positive and negative charges experience an attractive force, which should result in immediate movement and reasonably quick neutralization, limited by viscosity. However, the results showed that charge separation was well maintained for lengthy periods of time. Even if some barrier were formed between the respective charged zones, one might have expected a huge charge accumulation immediately across the barrier, as in a capacitor; however, the charges were reasonably well distributed over the full extent of the respective zones. Hence, it would appear that the separated charges are constrained in some fashion, as though confined in some kind of lattice, like n- and p-regions of semiconductors. Water appears able to adopt two structural networks that have mirror symmetry to one another. The fact that these networks are macro phenomena deserves further study.

A second and related issue is the potential for disturbance of these structural networks. It is now established that when water is left standing for long periods, it develops thixotropic properties, implying macrostructure.7 Such macrostructure is expected to be fragile. The fact that removing and inserting electrodes did not apparently ruin the charge-containing structure implies that, once formed, the structural network can re-form rather readily. This is an additional subject requiring further study.

Another significant issue is the correlation between pH difference and charge difference. Regions of low pH had excess positive charge, whereas regions of high pH had excess negative charge. If that were not the case, then the charge should not have been so easily recoverable. The question arises whether the correlation between pH and charge seen here might be more general. One might speculate, for example, whether ordinary acidic solutions, which have low pH, might contain net positive charge, while ordinary basic solutions might contain net negative charge.

Although the routinely invoked “law of electroneutrality” implies that net charge should not be possible in any volume of fluid, this “law” may not be inviolable. For example, clouds, essentially water, contain vast regions of net charge. In the present experiments as well, half-chambers with net charge could be sustained for long periods. Hence, the law of electroneutrality is not necessarily inviolable over restricted volumes, and the possibility that acidic/basic solutions might contain net charge should not be reflexively dismissed.

Hence, the results raise a number of questions that warrant further pursuit. At the very least, they open the possibility that ordinary water might have the capacity not only to store charge but to permit effective recovery of charge. Water may well be an unexpectedly effective charge-storage medium.


1. Fried I. The Chemistry of Electrode Processes. London: Academic Press; 1973.
2. Bard AJ, Faulkner LR. Electrochemical Methods: Fundamentals and Applications. New York: Wiley; 2001.
3. Klimov A, Pollack GH. Langnuir. 2007;23:11189–11895. [PubMed]
4. Macounova K, Cabrera CR, Holl MR, Yager P. Anal. Chem. 2000;72:3745–3751. [PubMed]
5. Cabrera CR, Finlayson B, Yager P. Anal. Chem. 2001;73:658–666. [PubMed]
6. Conway BE. Electrochemical Supercapacitors: Scientific Fundamentals and Technological Applications. London: Kluwer Academic Publishers; 1999.
7. Vybiral B. In: Water and the Cell. Pollack GH, Cameron I, Wheatley D, editors. New York: Springer; 2006. pp. 299–314.