Secondary Logo

Journal Logo

Gas Exchange

Photolytically Driven Generation of Dissolved Oxygen and Increased Oxyhemoglobin in Whole Blood

Monzyk, Bruce F.*; Burckle, Eric C.*; Carleton, Linda M.*; Busch, James*; Dasse, Kurt A.; Martin, Peter M.; Gilbert, Richard J.†§

Author Information
doi: 10.1097/01.mat.0000219086.39192.9a
  • Free


Despite the presence of reduced mortality in recent decades for most major diseases, the mortality rate associated with chronic lung disease has continued to rise. This is largely due to a lack of emerging therapies and inadequate ways to provide intermediate (“bridge”) or long-term respiratory support. Although numerous artificial lung technologies have been proposed and implemented,1–6 none has shown sufficient yield, safety, and ease of use to support broad deployment in the clinical arena. Furthermore, despite advances in lung procurement, preservation, and implantation,7–10 lung transplantation is inaccessible to most patients with respiratory failure because of the low yield of usable organs and the absence of methods to support patients pending transplantation.11

Conventional lung technologies are based on the delivery of oxygen to the bloodstream as gas through permeable hollow fibers.12–14 These systems depend on membrane diffusivity involving gas/liquid O2 and CO2 dissolution within pores and differential gas pressures to drive O2/CO2 exchange. The principal weakness of these systems is that they involve major diffusion boundary layers linked in series, which results in substantially slowed mass transport. They also require precise balancing of total gas pressure to enable simultaneous O2 and CO2 flux in opposite directions. As a result, these systems need a large surface area to achieve sufficient gas exchange. In addition, these systems require a continuous source of exogenous pressurized oxygen. We have circumvented these limitations by approaching the problem of intravascular oxygenation from a novel perspective. Rather than delivering oxygen to the blood (or back-flowing carbon dioxide against an influx of O2), we use photolytic energy to generate dissolved oxygen (DO) directly from the water already in the blood.

We have previously shown that it is feasible to generate DO from the water content of synthetic serum, based on the interaction of UVA light with a semiconducting TiO2 thin film.15 Our system thus mimics the optically driven oxidation reaction of photosynthesis (PS-II) found in cyanobacteria, higher plants, and algae. This reaction involves light absorption by metal chelate chromophores (chlorophylls) in a concerted reaction with a transition metal oxide cluster of manganese (Mn), resulting in charge separation–based O2 generation.16–19 The proposed photolytic lung technology builds on this known characteristic of the anatase form of TiO2 to serve as the chromophore yielding charge separation upon absorption of UVA light20–23 and uses it to oxidize water. These results have allowed us to propose a model for artificial respiration, which uses transition metal oxides that optically drive charge separation to carry out oxidation-reduction (“redox”) chemical conversions to generate oxygen from water present in whole blood (Figure 1).

Figure 1.
Figure 1.:
Diagram showing the fundamental mechanism of a photolytic DO generation and oxyhemoglobin formation. The photolytic technology is based on the ability of transition metal oxides to convert light energy to electric current; the resulting charge separation is used to generate oxygen from adjacent (non diffusion) water molecules supplied from plasma. The photochemical materials are designed to produce DO directly in an aqueous fluid (blood), without involving the gaseous phase, which is freely available for binding with hemoglobin. The photolytic artificial lung technology utilizes a semiconducting metal oxide material, TiO2, as both the photo-absorption element and photochemically driven chemical conversion center. The light energy produces charge separation (electron-hole pairs) in the semiconductor, which results in the production of active oxygen, hydrogen ions, and free electrons.

In the current project, we have demonstrated the feasibility of photolytic generation of DO and the ability to use this DO to form oxyhemoglobin in whole blood. We consider this a pivotal step in the development of an artificial technology for autonomous intravascular gas exchange.

Approach and Methods

Chemical Basis for the Photolytic Generation of Dissolved Oxygen

The following equations comprise the basis for the photolytic conversion of water to DO and the generation of carbon dioxide (CO2).

Reaction 1: Photolysis Yielding Charge Separation and Active Oxygen

where AO designates a solid-state active form of oxygen, for example, the peroxo “[TiO2(O2=)2+]” species occurring in the bulk solid phase of the photocatalyst film. The quotations indicate a surrogate formula for the transient photo-activated catalyst site within the TiO2 film, where the photon was absorbed (i.e., the “hole” or h+) or any locations within the solid to where the “hole” has migrated through electron exchange other than the surface; “scb” indicates that the electron produced on photon absorption is energetically transferred into the semiconductor band of the titania crystal. As shown below, AO has a very short life once it migrates to the surface of the photocatalyst that is in contact with the water supplied from the blood plasma. This migration step reforms the photon absorption bulk titania film site as follows.

Reaction 2: AO Migration to the Thin Film Surface and Hydration to Adsorbed Peroxy Species

The H2O present on the DO generating surface is supplied from the bulk aqueous phase (nominally 55.5 molar) and thus does not represent a significant diffusion boundary layer. Theoretically, water diffusion rate constraints would be expected only at very high lamp intensities and the highest DO flux values, a limitation not expected for the proposed technology based on current work. Once at the surface, DO is generated by spontaneous disproportionation and without involving a gaseous phase in the following manner.

Reaction 3: Disproportionation

where DO = dissolved oxygen, O2(aq). The hydrated surface titania species is regenerated at the same time where it is ready to undergo the next DO generation cycle. The DO diffuses out of the nanoporous surface at a flux proportional to the lamp intensity, the quantum yield, and the overall rate of reactions 1, 2, and 3. The DO migrates through the blood plasma, where it is taken up rapidly by red blood cells (RBC) and hemoglobin.

Reaction 4: Oxygenation of Hemoglobin

The hydrogen ions from Reaction 2 transfer through the aqueous phase by the well known “hopping” mechanism and react with the bicarbonate ion in the blood to enable the release of carbon dioxide (CO2) in the manner already involved in the natural lung.

Reaction 5: CO2 Removal Using Physiological Chemistry, Involving Protonation

This reaction is followed rapidly by:

Reaction 6: Spontaneous Catalyzed Dehydration

The chemical substrate for DO formation is a small amount of water derived from the blood. The formation of DO within the TiO2 ceramic nanoporosity prevents direct contact of blood cells to the DO formation region. The illumination region is only solid state and does not contact the aqueous or blood phases.

Fabrication of Titanium Dioxide Thin Films

The photoactive construct consists of a solid state layered structure, consisting of a transparent glass or quartz substrate, onto which a conducting film and then a photocatalyst film has been deposited, similar to that previously described.15 The thin film fabrication methodology was improved in the present work (see below) as was the method for O2 analysis.

Photocatalyst films were prepared as follows for testing in a batch apparatus. Glass substrates were 25 mm × 9 mm plates with 98% transmission at the desired wavelength. Thin (<100 nm) metallic (Ti) or semiconductor (indium tin oxide, ITO) conductive film(s) or grids were laid down on a glass surface using conventional vacuum sputter coating procedures. The photoactive layer consisted of a film of titanium dioxide (TiO2), deposited either by sputter coating or formed by sol-gel processing on top of the conducting film. The sol-gel method for preparing the photocatalyst consisted of the following procedures: The anatase TiO2 powder was HF acid treated before deposition to enhance adhesion by mixing 1 g of anatase TiO2 in 80 ml of a 1N HCl and 0.1% HF solution for 1 minute. The resulting slurry was divided equally into two centrifuge tubes and centrifuged until sedimentation was achieved. The acid was decanted and replaced with water and the particles resuspended. The samples were centrifuged and the liquid decanted. This water rinse was then repeated, and after the second water rinse and decanting, 40 ml of isopropanol (iPrOH) was added and the particles again re-suspended. Additional film coatings were performed by using manganese (IV) oxide, MnO2 (particle size <5 um, Aldrich Chemical Co, Milwaukee, WI) to enhance peroxide disproportionation rates.

Sol-gel oxide films were generated by using spin coating techniques. By this procedure, glass slides containing the conducting layer were placed on a vacuum chuck and rotated at 1,000 rpm. For the TiO2 coating, 0.5 g of the acid treated material was added to 40 ml iPrOH and mixed for 30 minutes; 0.050 ml H2O and 0.100 ml titanium (IV) tetra (isopropoxide) (TTIP), a sol-gel particle cross-linking reagent, was added to this solution. After mixing for 30 minutes, the solution was added drop-wise to the rotating substrate to a total volume of 12 mL. In the case of the constructs containing MnO2, after the addition of 9 mL of TiO2 slurry, 0.20 g MnO2 was added to the remaining slurry. Then 4 mL of the resulting solution was added to the substrate at spin coating conditions. Oxidation enhancing modifications of this technique were additionally examined, namely those where RuO2/Pt doped TiO2 (0.125g) was added to 10 mL iPrOH and added in place of the TiO2 slurry. After 15 minutes of mixing the above slurry, 50 μl water and 25 μl TTIP were added as sol gel binder for the anatase particles and allowed to mix for an additional 15 minutes. This solution was then added drop-wise to the substrate with spinning for a total volume of 9 ml. The sol-gel coated samples were all allowed to air dry at room temperature overnight then placed in a preheated tube furnace and heated at test temperature for 45 minutes under a 1 l/min flow of nitrogen.

Lamp Source

UVA light from an EFOS Lite unit was directed to the reaction chamber through a liquid light pipe after being filtered to produce light of only 365 nm. The light power at this wavelength was 88.1 mW/cm2 determined at the exit point of the light pipe with a Tamarack Model 157 hand-held traceable calibration photometer. Heating of the photocatalyst during illumination was minimized by the fact that TiO2 film efficiently absorbed light at 365 nm, thus resulting in little wasted light.

Batch Testing of Photocatalytic Chemistry

Photocatalyst films were prepared and assessed in a batch testing apparatus. Testing of parameters associated with photolytic production of DO from water was performed in the liquid phase cell, using a slurry or deposited film of the TiO2 material being tested immersed in a solution of ferric ions at pH 1.9 as an electron absorber The use of ferric ions allows the tests to be run within the setting of the electrically isolated cell, in which the ferric ions are maintained in solution by the low pH. The ferric ions (Fe3+aq) are converted to ferrous ions (Fe2+aq) during photolysis by chemical reduction by the photolytically mobilized electrons (escb) from the TiO2 photo catalyst reaction (Reaction 1 followed by

). As Fe2+aq ion is slow to oxidize in acidic media, the co-produced DO production rate can be measured directly with a Clark Cell and as the rate of ferrous ion formed. The DO is produced by Reactions 1, 2, and 3, where the escbis consumed through ferrous ion formation and is replaced within the TiO2 from adjacent water molecules with minimal diffusional constraints as the water concentration is exceedingly large (55 mol/l) adjacent to the photocatalyst surface. This oxidation to DO occurs despite the high thermodynamic stability of liquid water, due to the high energy level of the 365 nm photon. In the absence of ferric ions, there is either no net DO generation as the photo-generated electrons remain available to reduce DO back to water at a rate only a little slower than its formation. Therefore, for photocatalyst preparation for the flow cell, uniform film preparation was important and a bias voltage was applied to direct photo-produced electrons away from the aqueous interface at to the current collector film.

Flow-Through Test Cell for Assaying DO Generation and Hb Oxygenation

A flow-through, divided, photolytically driven electrochemical (PDEC) test cell was constructed to contact flowing blood with photolytically generated DO (Figure 2). This cell was a modified FM01-LC Electrolyser, operating in a divided cell mode with a Nafion™ cation exchange membrane. The anode was optically transparent and photons were supplied by side-on illumination using an EFOS Lite UVA light source filtered to 365 nm (see above). The uncoated glass or quartz side of the plates was illuminated side-on by filtered UVA light. The catholyte was Locke's-Ringer solution15 and the anolyte fresh whole bovine blood containing the anticoagulant, heparin sulfate. The blood was obtained from a local slaughterhouse for use on the day of the experiment, thus eliminating the need for extended preservation. A photograph and schematic of the test system are shown in Figure 3. Fluids were maintained at 37°C, using a glass in-line heat exchange jacket and flow was 80 ml/min by a Harvard peristaltic pump. Data collected were pH (glass electrode with calomel reference), electrical current (measured using a Fluka 87 Volt-Ohm meter, VOM, in μA or mA mode as required), temperature, DO, and oxyhemoglobin (O2Hb). The lamp intensity was varied with and without bias voltage. Control tests made during each run indicated that both bias voltage and UVA illumination was required for significant DO formation rate or electrical current flow to occur.

Figure 2.
Figure 2.:
Flow-through cell for DO generation and measurement. A, Schematic illustrating the individual elements of test cell construction. For purposes of clarity, film thicknesses are not drawn to scale. Shown are the configuration of the cell components, including photoactive films, light source, and inlet/outlet for both anolyte and catholyte. B, The flow-through test cell was constructed by using a base of a modified FM01-LC Electrolyser, operating in a divided cell mode with a Nafion™ cation exchange membrane. The key physical attributes of photolytic test cell includes: overall dimensions (height × width × length): 134 × 270 × 90 mm, active electrode area 0.0128 M2/0.0256 M2, current density 60 kA/M2.
Figure 3.
Figure 3.:
Test flow system and sensors employed in the assay photolytic oxygenation of blood. Shown is the test flow system depicting the relation of the photolytic reaction chamber to pumps and flow/oxygen sensors. Blood was maintained at 37°C, using an in-line heat exchange jacket and flow was 80 cc/min by a Harvard peristaltic pump. Data collected were pH, electrical current, dissolved oxygen, or DO, the partial pressure of oxygen, or pO2, the partial pressure of carbon dioxide, or pCO2, and oxyhemoglobin as a function of UVA irradiation (shown) and the presence or absence of bias voltage. A, Actual photo of system; B, schematic depiction of system and system components.

Measurement of Blood Gas Parameters

Conventional blood gas analyses were made by using a blood gas analyzer. Precise volumes of blood samples (250 μl) were withdrawn at specific times by a needle-fitted syringe from an in-line septum access port and analyzed using the NPT7 Blood-Gas Analyzer (Radiometer, Inc., Cleveland, OH). This instrument provided a readout of critical blood gas parameters including total hemoglobin (ctHb), percent hemoglobin as oxyhemoglobin [FO2Hb(%)], percent hemoglobin as deoxyhemoglobin [FHb(%)], percent hemoglobin as methemoglobin [FmetHb(%)], total viable hemoglobin concentration [sO2(%)], pH, and temperature. Since the goal of these experiments was to demonstrate changes in [FO2Hb(%)], pH, [sO2(%)] and pH as a function of UV light exposure of the photoactive film, measurements of the partial pressure of O2 and CO2 were not obtained. An in-line Clark Cell electrode was also used to monitor DO concentration in real time to insure contamination from the atmosphere was avoided and to provide an independent verification and measurement of this parameter. For the purposes of the concept validation experiments, a single blood flow rate and temperature was employed to minimize the number of potential variables while variables affecting photocatalyst film construction, analytical methods development, biocompatibility, and DO generation rate were assessed.

Real-Time Measurement of DO

A liquid phase in-line reaction chamber was used to monitor dissolved oxygen production in both the batch and flow-through test cells (see above). This device utilized a Clark Cell and electrode to measure DO through the selective diffusion of O2 through a porous hydrophobic membrane to an electrochemical sensor. A conventional two-point calibration procedure was used to calibrate the DO sensor. A liquid phase, air-tight reaction chamber was used to monitor real-time production of DO. Calibration points were 2.35 ml of air-saturated water at 36°C (217.2 nmol O2/ml) added to the cell before taking measurements, then emptied and flushed with nitrogen to obtain zero-oxygen. In the case of batch cell testing, Locke's-Ringer solution was added. In the case of the flow cell tests, whole blood flow initiated, after which the bias voltage was applied, generally 1 Volt. The UV light was cycled on and off to ensure the oxygen production was a result of a photo-catalytic phenomenon, not a galvanic or electrolytic one caused by the biased voltage.

Electron Microscopic Examination of Photoactive Thin Film Exposed to Whole Blood

Scanning electron microscopy (SEM) was used to assay the surface of the photocatalyst film before and after exposure to heparinized whole bovine blood for potential biofouling. Each sample was affixed to standard holders for examination with a JEOL 840A SEM. Electron images were obtained at multiple magnifications with visualization of features to better than 1 micron resolution. An Oxford INCA 300 Energy Dispersive Spectrometer (EDS) was used to determine the elemental composition of the films before and after exposure to whole blood.


Effect of Titanium Dioxide Thin Film Properties on Photolytic DO Generation

Experiments (n = 40 experiments) were performed to determine conditions for a significant photocatalytic effect on DO production, reduction of O2/electron recombination, application of bias voltage, determination of sealing viability from atmospheric contamination, and drift.

In range-finding testing, DO generation was achieved in the liquid phase cell by using a slurry of TiO2 and ferric ion (see above). These tests confirmed that photo-activated TiO2 can be made to replace photo-ejected electrons from water molecules, thereby producing DO. Under control conditions, with TiO2 suspended in Locke's Ringer and light delivered by a mercury lamp, but without ferric ions, no DO generation was measured (Figure 4A). When TiO2 was suspended in 9.98 mmol/l H2SO4 solution containing 3.96 mmol/l FeCl3, a steady-state increase of DO was observed after initiation and continuation of light exposure (Figure 4B). Note that in the absence of flow, DO concentration increases to a steady-state maximum. When TiO2 was suspended in 9.98 mmol/l H2SO4 solution containing FeCl3 with the light cycled on and off a number of times, an explicit dependence on photolytic energy for DO production was demonstrated (Figure 4C). When the lamp was turned off, oxygen production ceased and oxygen concentration decreased.

Figure 4.
Figure 4.:
Demonstration of photolytic DO generation in batch cell preparation (non-flowing). A, Control condition, TiO2 (25.0 mg) suspended in 1.5 ml Locke's Ringer (7.2 pH) at 36°C with light delivered by mercury lamp (0.93 mW/cm2 at 365 nm) without ferric ions, resulting in no DO measured. B, TiO2 (16.0 mg) suspended in 2.0 ml of 9.98 mmol/l H2SO4 solution containing 3.96 mmol/l FeCl3 (1.9 pH) at 36°C with light delivered by mercury lamp (0.93 mW/cm2 at 365 nm), demonstrating increase of study state DO. C, TiO2 (16.0 mg) suspended in 2.0 ml of 9.98 mmol/l H2SO4 solution containing 3.96 mmol/l FeCl3 (1.9 pH) at 36°C with light delivered by UVA lamp (88.9 mW/cm2 at 365 nm). Lamp was cycled on and off a number of times as indicated, demonstrating that photolytic energy is critical for DO production and that DO is not formed without photolysis.

Samples with varying thicknesses of titanium metal (conducting layer) and TiO2 (photoactive layer) were obtained, and DO generation in response to UV light was measured (Table 1). In general, at a film thickness of 1,200 angstroms, there was a significant reduction in resistance and an increase in DO generation. Similarly, we tested various metal materials of different thicknesses for conduction, including nickel, chromium, and ITO (Table 2), and found that ITO at approximately 1,300 angstroms offered the best yield in terms of DO generation rate. This loss/gain of conductivity of the thin layers was significant, since this property affects their ability to act as electron sinks and thus avoid electron recombination. It should be pointed out that ITO was used because it is a transparent, conductive material that will allow illumination of the photoactive layer by passing UV light through the substrate base, rather than through the solution. This method of illumination is important as it prevents direct interaction of the solution medium with UV radiation.

Table 1
Table 1:
Samples with titanium metal of various thickness as the conducting layer and TiO2 as the photoactive layer
Table 2
Table 2:
Samples with Ni, Cr, and ITO of various thicknesses as the conducting layer and TiO2 as the photoactive layer

We further tested the effect of an applied bias voltage on the rate of oxygen generation (Figure 5). When +1V was applied to the system, the maximum rate obtained was 123.3 nmol/ml per hour. This increased to 171.4 nmol/ml per hour when +2 V was applied. Both UV light and bias voltage were required for oxygen generation, and direct electrolysis of water was not observed even at the +2V bias voltage level. The presence of the +1V potential between the conducting layer of the construct and a Pt counter electrode during illumination produces an electric field across the photolytic layer, which directs electron flow from the TiO2 layer to the Ti conductor film. When the voltage bias is removed and UV illumination ceases, the DO concentration decreased indicating that electrons were available on the photocatalyst surface to drive the reformation of water from the DO. As this decrease was not observed for sputter-coated TiO2 films, it was concluded that the sol-gel TiO2 crystallites do not have the long range order exhibited by sputtered coatings.

Figure 5.
Figure 5.:
Dependence of DO generation on the application of a bias voltage. Real-time demonstration of DO production in a glass substrate (9 mm × 25 mm) containing 1,230 A˚ Ti and TiO2 in 2.35 ml of Locke's solution (7.2 pH) at 36°C. Bias voltage of +1 V was applied by using photoactive construct as the anode and a Pt wire as the cathode, and UV light supplied. Bias voltage of +1 to +2 V was applied using photoactive construct as the anode and a Pt wire as the cathode. UV light was supplied by the EFOS lamp.

Photolytic Oxyhemoglobin Generation

Since the photocatalyst film constructs used in these experiments were considered early in development, we did not attempt quantification of the rate of oxyhemoglobin formation but rather focused on confirming the phenomenon of photolytic DO generation and oxyhemoglobin formation. Filtered UV illumination of the metal oxide light–absorbing surface opposite the blood in the flow-through cell described above resulted in near complete oxygen saturation of the blood circulating on the opposite side of the photolyzed surface (Figure 6). These observations were characteristic of four separate experiments performed under similar conditions using several different sputter-coated photocatalyst preparations, and reflect the chemical effects depicted in Reactions 1 to 4. In the experiment shown, the fraction oxyhemoglobin increased from 83% to 92% and remained stable throughout the trial period. The fact that the FO2Hb and SO2 curves varied in a parallel manner with SO2 higher than FO2Hb demonstrates that near-complete oxygenation of hemoglobin content of the blood was achieved. It should be noted that the viable, nonviable, and total Hb levels remained constant during the complete period of blood circulation through the flow cell.

Figure 6.
Figure 6.:
Photolytic generation of DO and enhanced oxyhemoglobin in whole blood. The test apparatus shown in Figures 3 and 4 was used to determined DO and oxyhemoglobin formation as a function of light exposure. The catholyte was Locke's-Ringer solution, and the anolyte was bovine blended venous/arterial whole blood anticoagulated with heparin and thermostated at 37°C. Flow was recirculated by using by a peristaltic pump. The fraction (F) of oxyhemoglobin (O2Hb), denoted FO2Hb expressed as a percentage of total Hb present, rapidly increased with illumination from 83% to 92% and then remained stable throughout the trial. The term FO2Hb indicates the percentage of hemoglobin, which contains bound oxygen, i.e., hemoglobin saturation. The term SO2 indicates the fraction of O2 saturation expressed as a percentage of the maximum solubility of O2 in whole blood at test temperature.

The effect of UV light illumination of the photocatalyst on the pH, plasma saturation of oxygen, hemoglobin saturation of oxygen, and the percentage of blood methemoglobin were determined at various times (16 to 460 minutes) in a recirculating flow loop while collecting and assaying 11 samples (Table 3). The observation that FO2Hb did not reach maximum saturation, despite recirculation of the blood through the DO generating apparatus, may be explained by the system reaching a steady state with competition between electron removal by conduction to the ITO collector and the recombination of these electrons with DO. Hence, as the DO reaches a high level the back reaction to form water will also increase in rate if residual electrons are available from the photocatalyst (at so called “shallow” sites in the semiconductor band of the TiO2). The fact that the pH of the anolyte dropped during the initial formation of oxyhemoglobin can be understood on the basis of conventional blood chemistry coupled with the previously described overall photochemistry (Reactions 1 through 4) as follows:

Table 3
Table 3:
Blood oxygen analyses

The electrons generated provide the basis for the observed electrical current, and the H+ ions caused the observed small drop in pH. The drop in pH is small as most of the H+ ions were immediately consumed through CO2 formation from the blood's bicarbonate buffer in the normal physiologic manner (Reactions 5 and 6), resulting in the stabilization of blood pH and enabling the release of CO2 from the blood. We considered that the exposure of the TiO2 surface to continuous UV light (even without direct exposure of the blood) could have toxic oxidant effects,24,25 resulting in an increased percent of nonviable hemoglobin, especially methemoglobin. However, the percentage of methemoglobin measured was stable and low throughout the trial, in the range of 2% to 3%, indicating the absence of blood degradation detection by this method.

Since biofouling of the DO generation surface by blood cells could affect the rate of DO generation, we assayed the surface appearance with SEM after UV irradiation of the photocatalyst while the DO generation side of the construct was exposed to heparinized whole blood for 30 minutes. Figure 7 demonstrates that this treatment of the TiO2 construct film yields fundamentally clean surfaces, containing only a few small adherent inorganic crystalline residues from the vacuum sputter coating preparation of the original TiO2 film and no blood cells or clots.

Figure 7.
Figure 7.:
Scanning electron microscopy to assay for surface biofouling. A and B, SEM of thin film surface after UV light exposure but before contact with whole blood at x100 (A) and x2,500 (B) magnifications. C and D, SEM of thin film surface under the same conditions after contact with whole bovine blood at x100 (C) and x2,500 (D) magnifications. EDS spectra of observed surface grains indicated inorganic residues only in B and buffer salts in D.


Photolytic reactions are ubiquitous mechanisms in nature through which light energy is used to drive metabolism. In this study, we applied photolytic principles to the development of a novel respiratory device that generates increased intravascular oxygen and augments oxyhemoglobin formation in blood without the physical delivery of oxygen gas or use of gas cylinders. During photosynthesis, oxygen is generated from water, using photolytic energy under mild conditions of pressure, temperature, and pH, while releasing hydrogen ions for driving ion pumps and forming ATP, and energetic electrons to drive a cascade of redox reactions (“the electron transport chain”) to drive additional energy production. In our system, a semiconducting metal oxide film was designed as the photo-absorption element, the anatase form of titania, TiO2.26–28 Photolysis of this ceramic oxide results in useful charge separation, and which, on further development, should yield oxygen production for much longer durations than achieved by organic chemical-based photosynthetic pigments (i.e., the chlorophylls). Importantly, the light energy associated with activation by a 365 nm UV lamp or 354 nm laser light selectively excites the ligand-to-metal electronic transition of the TiO2 semiconductor (350- to 389-nm band, or at least 3.2 eV), thus resulting in minimal wasted radiation or transmission. Future work could extend this useful wavelength region further by using metal oxide blends with dopants and sensitizers, which will aid in reducing the size and power demands.

In our initial work,15 we demonstrated that photoactivation of the proposed thin film in contact with surrogate blood serum induced the following reactions in significant yield: (1) UV light is absorbed into the TiO2 film producing electrical charge separation which can be detected using appropriate dyes for the “hole” and conductance band electron. (2) Charge separation drives the conversion of water, present in large excess from the plasma, to form the “active oxygen” (AO) intermediate in the nanoporous metal oxide film. (3) Similar to photosynthesis, AO spontaneously disproportionates into oxygen as DO without forming a gaseous phase as a first step, thereby eliminating a key diffusion boundary layer. (4) DO diffuses into the aqueous phase (surrogate blood serum). (5) Simultaneously, to avoid recombination with AO, the freed electrons are conducted away to the cathode as evidenced by flow of electrical current. (6) The freed H+ rapidly combines with bicarbonate ions in the plasma by a protonation reaction, yielding free carbon dioxide. It may be advantageous in the future to combine these hydrogen ions with the bicarbonate ions and freed electrons at the cathode to form organic compounds that may be easily eliminated through the urine or GI tract. In either case, the H+ ions are not required to diffuse from the point of formation but use the well-known proton “hopping” mechanism,29 thus eliminating another diffusion boundary layer. In lieu of CO2 “fixation,” we project that CO2 (aq) may be eliminated by per-evaporation with a CO2 gas selective biocompatible membrane using the high [CO2 (aq)] that will have developed across the photolytic cell.

In the current study, we have shown that photolytically driven mechanisms can be used to generate sufficient DO (as a function of unit area and recirculation time) to yield a measurable increase of oxyhemoglobin in flowing whole blood. To determine how far the technology needs to evolve in DO production rate, we performed the following calculation. Assuming that one molecule of O2 would be formed for every four electrons produced (assumed reaction stoichiometry), we used the measured electrical current to calculate that the maximum rate of oxygen generation with the current apparatus is 1.08 ml O2 per m2/min. Projecting this result to an alveolar surface area of 75 m2, as is the case for the human lung, the output of this system would be approximately 81.0 ml O2/min. This value is a significant fraction of the net oxygen flux traversing the normally functioning pulmonary capillary membrane, that is, 250 ml O2/min. Since the slow process of liquid dissolution of oxygen gas and permeation of the resultant DO through the alveolar and pulmonary vessel wall membranes are not required for the technology under development, a major impediment faced by current artificial lung technologies is avoided. In addition, the requirement for counterflow of O2 and CO2 is required for both the incoming air and exhaled CO2 in normal lung is avoided by separating the removal steps for these two components. Instead, because DO flux is not dependent on diffusion across a gas/liquid boundary or against an outflow of diffusing CO2, a high DO flux entering the blood is conceivable. The TiO2 semiconductor material is considerably more durable than PS pigments (i.e., naturally occurring chlorophyll) and, importantly, is selectively excited by light of a narrow UVA bandwidth, thus minimizing heat production and avoiding the potential exposure of blood cells to UVA light. The catalyst film, which converts AO into DO, is located within the photoactive film and buried within the nanoporosity so that in the end, only DO exits the surface pores and is exposed to blood.

Although the current results provide proof of concept for the photolytic oxygenation of whole blood, there are several important questions which need to be resolved if this approach is to constitute a viable artificial lung technology: (1) Improved photolytic yield. The enhancement of quantum efficiency of DO generation will allow the further reduction of device size, device complexity, and photolytic power requirements. This can be accomplished by modification of nanocrystal structure design and uniformity of film deposition, improved photoconductivity and enhanced photoelectron removal (and thus reduced electron recombination). (2) Integration of photolytic DO generation with blood flow at a microfluidic scale: The current results assume pragmatic dimensions for the photolytic volume. However, to maximize blood oxygenation, microfluidic devices are needed which incorporate an architecture promoting extended contact between the blood cells and the gas exchange surface. (3) Hemocompatibility: Although the absence of hemolysis or methemoglobin generation is encouraging, there remains concern whether long-term photolytic DO generation and its byproducts are toxic in terms of blood cell viability and mutagenicity. Enhancements of the technology along these lines should provide more substantial rationale for the development of such autonomous blood oxygenation technologies.

In summary, we have demonstrated the feasibility of a device, which generates dissolved oxygen through photoinduction, thus allowing auto-oxygenation of whole blood through its own water content. We anticipate that the photolytic lung technology will compete favorably against conventional ventilators since the photolytic technology achieves intravascular gas exchange without an exogenous source of gas at mild conditions, and can be miniaturized. The photolytic artificial lung technology may be considered as a subset of a broader technology platform, which brings together several physical systems in close proximity of each other to obtain synergistic chemical changes and useful component. These systems include a sensitive and complex aqueous phase, photolytic energy to provide “charge separation,” thus simultaneously yielding chemical and electrical energy for photolytically driven anodic and cathodic chemical reactions. The achievement of nanoscale photoinduction to power key chemistries is expected to comprise a broad platform from which numerous biomedical technologies may ensue.


1.Bagley B, Bagley A, Henrie J, et al: Quantitative gas transfer into and out of circulating venous blood by means of an intravenacaval oxygenator. ASAIO Trans 37: M413-415, 1991.
2.Funakubo A, Higami T, Sakuma I, et al: Development of a membrane oxygenator for ECMO using a novel fine silicone hollow fiber. ASAIO J 42: M837-840, 1996.
3.Kitano Y, Takata M, Miyasaka K, et al: Evaluation of an extracorporeal membrane oxygenation system using a nonporous membrane oxygenator and a new method for heparin coating. J Pediatr Surg 32: 691-697, 1997.
4.Moreno-Cabral RJ, Dembitsky WP, Adamson RM, Daily PO: Percutaneous extracorporeal membrane oxygenation. Adv Card Surg 5: 163-179, 1994.
5.Mortensen JD, Berry G: Conceptual and design features of a practical, clinically effective intravenous mechanical blood oxygen/carbon dioxide exchange device (IVOX). Int J Artif Organs 12: 384-389, 1989.
6.Stammers AH, Fristoe LW, Alonso A, et al: Clinical evaluation of a new generation membrane oxygenator: a prospective randomized study. Perfusion 13: 165-175, 1998.
7.Date H: Current status and future of lung transplantation. Intern Med 40: 87-95, 2001.
8.Harringer W, Wiebe K, Struber M, et al: Lung transplantation 10-year experience. Eur J Cardiothorac Surg 16: 546-554, 1999.
9.Hosenpud JD, Bennett LE, Keck BM, et al: The Registry of the International Society for Heart and Lung Transplantation: fifteenth official report 1998. J Heart Lung Transplant 17: 656-668, 1998.
10.Mentzer SJ, Swanson SJ: Lung transplantation, in Austen KF, Burakoff SJ, Rosen FS, Strom TB (ed), Therapeutic Immunology, Blackwell Science Inc., Walden MS, 2001.
11.De Meester J, Smits JM, Persijn GG, Haverich A: Listing for lung transplantation: life expectancy and transplant effect, stratified by type of end-stage lung disease, the Eurotransplant experience. J Heart Lung Transplant 20: 518-524, 2001.
12.Federspiel WJ, Hewitt TJ, Hattler BG: Experimental evaluation of a model for oxygen exchange in a pulsating intravascular artificial lung. Ann Biomed Eng 28: 160-167, 2000.
13.Golob JF, Federspiel WJ, Merrill TL, et al: Acute in vivo testing of an intravascular respiratory support catheter. ASAIO J 47: 432-437, 2001.
14.Zwischenberger JB, Anderson CM, Cook KE, et al: Development of an implantable artificial lung: challenges and progress. ASAIO J 47: 316-320, 2001.
15.Dasse KA, Monzyk BF, Burckle EC, et al: Development of a photolytic artificial lung: preliminary concept validation. ASAIO J 49: 556-563, 2003.
16.Limburg J, Vrettos JS, Liable-Sands LM, et al: A functional model for O-O bond formation by the O2-evolving complex in photosystem II. Science 283: 1524-1527, 1999.
17.Vrettos JS, Brudvig GW: Water oxidation chemistry of photosystem II. Philos Trans R Soc Lond B Biol Sci 357: 1395-1405; discussion 1419-1320, 2002.
18.Yachandra VK, DeRose VJ, Latimer MJ, et al: Where plants make oxygen: a structural model for the photosynthetic oxygen-evolving manganese cluster. Science 260: 675-679, 1993.
19.Yachandra VK, Sauer K, Klein MP: Manganese cluster in photosynthesis: Where plants oxidize water to dioxygen. Chem Rev 96: 2927-2950, 1996.
20.Fernandez-Ibanez P, Blanco J, Malato S, de las Nieves FJ: Application of the colloidal stability of TiO2 particles for recovery and reuse in solar photocatalysis. Water Res 37: 3180-3188, 2003.
21.Hagfeldt A, Gratzel M: Molecular photovoltaics. Acc Chem Res 33: 269-277, 2000.
22.Topoglidis E, Campbell CJ, Palomares E, Durrant JR: Photoelectrochemical study of Zn cytochrome-c immobilised on a nanoporous metal oxide electrode. Chem Commun (Camb) 1518-1519, 2002.
23.Tsai P, Wu CT, Lee CS: Electrokinetic studies of inorganic coated capillaries. J Chromatogr B Biomed Appl 657: 285-290, 1994.
24.Beutler E: Methemoglobinemia and other causes of cyanosis, in Beutler E, Lichtman MA, Coller BW, Kipps TS (ed), Hematology, New York, McGraw-Hill, 1994, 654-662.
25.Gibson QH: The reduction of methemoglobin in red blood cells and studies on the cause of idiopathic methemoglobinemia. Biochem J 42: 13-23, 1948.
26.Asahi R, Morikawa T, Ohwaki T, et al: Visible-light photocatalysis in nitrogen-doped titanium oxides. Science 293: 269-271, 2001.
27.Chen F, Xie Y, Zhao J, Lu G: Photocatalytic degradation of dyes on a magnetically separated photocatalyst under visible and UV irradiation. Chemosphere 44: 1159-1168, 2001.
28.Garcia Rodenas LA, Weisz AD, Magaz GE, Blesa MA: Effect of light on the electrokinetic behavior of TiO(2) particles in contact with Cr(VI) aqueous solutions. J Colloid Interface Sci 230: 181-185, 2000.
29.Voth GA: The computer simulation of proton transport in biomolecular systems. Front Biosci 8: s1384-1379, 2003.
Copyright © 2006 by the American Society for Artificial Internal Organs