colloidal coal in water suspensions

Home Publishing Journals Energy & Environmental Science Advance Articles DOI: 10.1039/b923601p

Energy Environ. Sci., 2010 DOI: 10.1039/b923601p Paper

Energy & Environmental Science

Colloidal coal in water suspensions

Gustavo A. Núñez *a, María I. Briceño a, Daniel D. Joseph abc and Takeshi Asa aaNano Dispersions Technology, Inc., Bldg. 231, City of Knowledge, Clayton, Panama. E-mail:[email protected]; Tel: +507 64 00 64 09bDepartment of Aerospace Engineering and Mechanics, University of Minnesota, Minneapolis,Minnesota 55455, USAcDepartment of Mechanical and Aerospace Engineering, University of California, Irvine,California 92617, USA

Received 10th November 2009 , Accepted 8th January 2010

First published on the web 24th March 2010

We discuss the possible clean coal applications of colloidal suspensions of coal in water (CCW)manufactured with a proprietary wet-comminution device. These suspensions are a new materialwith new properties. First, the colloidal fraction plus water is a pseudo fluid good for transport,handling and suspension of large particles. Second, the surface area per unit volume of coalavailable for chemical reaction and burning is greatly increased and finally, CCW may be milledwith a third fluid, seeding the mixture with submicron coal.The colloidal nature of the majority ofparticles provides for very good features such as outstanding long-term stability, in contrast toregular coal water slurries (CWS) which rapidly sediment under storage. Moreover, the very smallparticles create an increased reactivity to combustion because small particles with large surface areareact faster than large particles with the same volume.

Broader context

We discuss the possible clean coal applications of colloidal suspensions of coal or pet-coke inwater (CCW) that were manufactured with a wet-comminution device. These suspensions are anew material with new properties. First, the colloidal fraction plus water is a pseudo fluid goodfor transport, handling and suspension of large particles. Second, the surface area per unitvolume of coal available for chemical reaction and burning is greatly increased and finally,CCW may be milled with a third fluid, seeding the mixture with submicron coal. Usingabundant raw sources, this material opens the possibility for producing homogeneous liquid

Colloidal coal in water suspensions (DOI: 10.1039/b923601p) http://www.rsc.org/delivery/_ArticleLinking/DisplayHTMLArti...

1 of 25 4/9/10 3:17 PM

fuels for power generation with more efficient and cleaner schemes, such as those involvinglow-speed diesel engines, gas turbines and gasification. Irrespective of the route taken toconstruct a greener economy for the world, abundant and evenly distributed fossil fuels such ascoal will be used in the foreseeable future. Utilizing these resources in CCW form maycontribute to their use with much less environmental impact by paving the road to lowemission electricity.

1.0 Introduction

In this paper we discuss the possible applications of colloidal dispersions of coal in water (hereaftercalled CCW), which we are able to manufacture at relatively low cost and in large quantities with aproprietary wet-communition device. These dispersions can be considered as a new materialbecause the coal particles do not settle but are held in suspension by Brownian motions. Coal–waterslurries used previously are dispersions of micronized coal with mean particle sizes greater thanfifteen microns; these dispersions are not colloidal (or the colloidal population is small) because theparticles settle rapidly under gravity leaving clear water behind.

A new material like CCW has a possibly vast field of applications in improvements andstrategies for new clean coal technologies. Because coal is cheap and abundant (per million BTU,compared with natural gas and oil), its global consumption will likely increase under anyforeseeable scenario as described in Ansolabehere et al.1 Its low cost and wide availability alsomakes it particularly attractive in major developing economies to meet their energy needs.However, the combustion of coal introduces pollutants harmful to the environment; the reduction ofharmful pollutants in the combustion of coal, especially CO2, is the goal of clean coal technologies.Our study targets the possible applications of CCW in these technologies.

At present, many efforts are being focused in reducing CO2 emissions and there are sometechnologies based on coal combustion or gasification, already on a commercial level, that manageto reduce the flue gas flow rates, thereby facilitating CO2 capturing and sequestration (CCS). Thesetechnologies are integrated coal gasification combined cycle (IGCC), oxy-fuel combustion and ultrasuper critical pulverized combustion with CO2 capture (USC) as also discussed by Ansolabehere etal.1 Whatever the route taken, CCS comes at significant cost and there is the remaining concernabout the long-term stability of the sequestered volumes in deep-ocean and underground reservoirs.

An alternative route to CO2 management has recently been advanced by Wibberley et al.,2 whoproposes to reduce the amount of CO2 emitted by a step improvement in the thermal efficiency ofdelivered electricity and using CCS as a last step to meet regulation targets. The improvement inthermal efficiency can be achieved by the adoption of low speed diesel engines (LSD) andcombined cycle gas turbines (CCGT) electricity generation schemes.

There is substantial prior work to substitute diesel fuel and heavy fuel oil for coal in dieselengines, especially for locomotives and small generation plants, but never with CO2 emissioncontrol as the main driver. In the case of gas turbines, there have also been many attempts to usecoal-based fuels, especially during periods of high prices in natural gas. The main barriers with thisapplication have been the very severe milling and cleaning that is required to reduce turbine bladeswear by coal particles and mineral matter.


2 of 25 4/9/10 3:17 PM

One possible route to fire coal in LSD and gas turbines is the development of ultra clean coalwater fuels (Wibberley et al.2). Historically, these fuels have been mostly targeted as substitutes forcrude oil with market penetration dependent on oil prices and oil price projections. Despite progressin milling technology, these fuels have typically had handling difficulties due to storage and evenin-line settling of particles. Ultra clean coal water fuels could also benefit from recent developmentsin coal cleaning, both chemical and physical (see Thambimuthu3 and Simeon4) but problemsassociated with large coal particles have to be addressed.

In the work that follows, we describe the manufacturing, physical characterization andcombustion of colloidal suspensions of coal in water. These suspensions contain a large population,close to 50% w/w, of submicron-sized particles. Because of this feature, a suspension having 60%of an Eastern bituminous coal exhibits outstanding stability to sedimentation and a flow behaviorthat makes it easy to handle under pumping conditions. Further, the suspension was tested as areburn fuel and it exhibited a notable performance in NOx reduction, much higher than drybituminous coal and, under some conditions, close to natural gas performance. To this end, wecontracted GE-Energy to perform pilot scale reburning tests to define the NOx reduction and carbonburnout with this suspension in their Santa Ana facility. In GE's experience, the best way to runsuch tests is to comparatively evaluate a baseline fuel and a test fuel, thereby highlighting thepotential benefits of the test fuel. Therefore, for the present work GE performed reburning testswith both a conventional coal grind and the colloidal coal suspension of the same coal type. NOxreduction and carbon burnout were measured as a function of key reburning test parameters.

While at first glance this could be considered a conventional coal water fuel, we assert that thiscoal suspension can be considered a new material. The macro-view of the fuel is that of a uniform,non-particulate liquid. Flow can be controlled with great precision (this is also a plus inapplications such as gasification) and it can be expected that the time scale of particle combustionbe reduced, along with an increase in particle reactivity (Davies et al.5). Combustion patterns maybe modified in the sense that particles larger than the submicron domain burn more efficientlybecause they are surrounded by a space heavily populated by very small, very reactive burningparticles (this is still a research issue that will be addressed later).

The foregoing discussion emphasizes the potential application of CCW in LSD and CCGT, withthe idea of reducing net CO2 emissions per unit of electricity generated. CCW may also address theissue of wear in cylinder walls of diesel engines and turbine blades due to large coal particles(Wibberley et al.2). Other applications in mind are in coal cleaning processes, as a boiler orreburning fuel, in gasification and oxy-combustion, in firing low-NOx burners, and in fuel cells.

Regarding coal cleaning, the process described in this work mills coal to colloidal sizes bycomminution, in the presence of water. This process can help improve mineral separation andenhance both physical and chemical cleaning procedures that are sensitive to particle size. Frothflotation and ash removal by chemical means are examples of this type of operation where reactionrates and efficiencies, as well as flotation, can be influenced by size.

As will be further described later, this material is different from other types of finely milled coal,such as micronized coal (DOE-Report NETL6). The latter has no significant submicron particles, asopposed to the material advanced here. Further, a colloidal suspension of coal in water can addressthe issues associated with firing coal in LSD and CCGT, especially concerning wear on cylinderwalls of diesel engines and turbine blades.

2.0 Experimental method


3 of 25 4/9/10 3:17 PM

2.1 CCW suspension preparation and properties characterization

The colloidal dispersions are prepared in two stages: first by a bench mill and then by ourwet-comminution device. The bench mill was manufactured by IKA®-Group. After grinding,samples were sieved using mesh size sieves 40 (400 m), 70 (212 m) or 140 (106 m), the passingparticles were retained and used to prepare coal suspensions with various water contents (30 to50%), surfactants and other types of additives. These mesh sizes are not foreign to coal-fired powerplants. In section 3.6 we describe the use of a 60% CCW, prepared in this fashion, as a reburningfuel.

It is noteworthy that a preliminary formulation study is first necessary in order to determine thetype and concentration of additives that are best suited to improve coal particle wetting and reduceviscosity. The additives were mostly surfactants and viscosity controlling agents, every type of coaltested usually required a specific formulation. In general, it was found that nonionic surfactantswere good wetting agents, in concentrations varying from 0.1 to 0.6 w/w %. Some of the additivesused to reduce viscosity by decreasing particle interactions, before or after the wet-comminutionprocess, were amines. The suspension formulation previous to the wet-comminution instance wasvery simple since what was basically required was a good wetting agent or a combination of twowetting agents. The idea was to have a uniform mixture with as low viscosity as possible.

Particle size of coal samples was determined by direct observation in an optical microscope, orby sieving using five or six different sieves ranging from 20 to 400 m, or using a laser diffractionapparatus made by Microtrac Corporation, Nanotrac model, with a measurement range from 8 nmto 6.5 m. None of these methods was sufficient to obtain a complete characterization of theparticle size distributions, but a combination of the three allowed for a good assessment of whatwas really in the suspension, before and after the wet-comminution process.

In our study, the percentage mass passing the 635 mesh size sieve (< 20 m) was used as anindicator of wet-comminution process efficiency (generation of colloidal particles), given thatmicroscopic observations generally showed that particles between 8 to 20 m were very scarce.

The preparation of the colloidal suspension of coal was centered in a technology that is totallybased on fluid mechanics principles. As mentioned above, a preliminary suspension was preparedin a tank with low agitation and the appropriate water and surfactant contents. This suspension isthen fed into a device that spins a film of fluid onto the walls of a cylindrical vessel at very highspeed and under cavitation free conditions. The resulting flow field induces a particle trap regionwhere coal particles are locally concentrated above their nominal value and under very high shear.Particles are then milled to very small sizes by a wet-comminution mechanism. Friction heating iscontrolled by a chilled water jacket around the vessel. A schematic view of the set up is shown inFig. 1.


4 of 25 4/9/10 3:17 PM

Fig. 1 Experimental set-up for the preparation ofcolloidal coal suspensions.

The energy consumed by the wet-comminution device was evaluated by monitoring the power(voltage and amperage) during the process. The latter has two components, the power required todrive the motor shaft and mechanical seal, and the net power consumed by the fluid duringcomminution. It was found that the net power divided by the mass flow rate, in terms of kW h ton−1,depended on coal content and viscosity of the preliminary slurry, exhibiting values of 30 to 80 kWh/ton. The energy consumed by the motor shaft and seal would account for 50 to 80% of the totalpower consumed.

Using the method described above, 100 gallons of CCW were prepared, using an Easternbituminous coal that was previously ground to 200 mesh. Several properties of this sample werecharacterized, namely, stability to sedimentation, shear viscosity and some of its combustionproperties when used as a reburn fuel.

The long-term stability of the sample just mentioned, prepared with the wet-comminutionmethod, was followed by observation of sample homogeneity. This included detection of clarifiedwater at the top and formation of a sediment layer. Clarification was evaluated by means of laserback-scattering scans. The apparatus used for this purpose was a Turbiscan Model MA2000 madeby Formulaction Corporation. A special glass vial containing the suspension sample was introducedin a chamber; a laser beam swept the vial from bottom to top. As it moved upwards, the lasertravelled along the sample up to the free surface. Receiving sensors collected back-scattered light,measured as a percentage of all the light intensity emitted.

To detect the formation of a sediment layer, a spatula was introduced in the sample, in a slowand vertical movement (similar experiments have been carried out by Boylu et al.7), up to thebottom of the jar containing the sample. The latter test only determined if a sediment layer wasthere, and no detection of sediment thickness or dispersability was intended. The same kind of testswas made for the samples in the special glass vials of the laser apparatus. Since these vials werenarrow, a thin wire was then introduced to check for sediment formation.

The shear viscosity of a conventional slurry and CCW was evaluated using a rotationalrheometer made by Rheometrics, model SR-2000, equipped with parallel plates. Both samples hadthe same coal content (60% Eastern bituminous coal) and the same amount of additives. A gravitydriven capillary viscometer was also used to evaluate flow properties of polymodal, 68.5% coalsuspensions having particles too large (>150 m) to allow for the use of a rheometer. This deviceconsisted of a 150 mL cylindrical container with a flush mounted glass, acrylic or PVC capillary inthe bottom center. Several capillary lengths (33 to 65 cm) and diameters (5 to 12 mm) were used fora complete evaluation. The experimental procedure was as follows. The container and capillary


5 of 25 4/9/10 3:17 PM

were filled with sample, while blocking the capillary outlet. Then the outlet was opened and thesample was allowed to run down to a receiving vessel while flow time was monitored by means ofa chronometer. The amount of sample collected was later weighed.

Shear stress for the capillary viscometer was calculated as

(1)

where d is the capillary diameter, s is the suspension density, H is the height of sample in thecontainer, L is the capillary length, g is the gravity acceleration. It has to be noted that changes in Hduring a test were kept very small relative to L ( H L), the capillary length.

The apparent shear rate, , was calculated as

(2)

where m is the sample mass collected during time t. The apparent viscosity was then calculated asthe ratio of shear stress over shear rate.

2.2 Combustion of CCW

2.2.1 Flame life time. Comparative flame time measurements were made in the laboratory byplacing samples of coal on aluminium dishes that had many small perforations to allow for aircirculation. Two kinds of samples were evaluated, colloidal coal and coal with regular grind. Thecoal used for these test was Guasare coal from Venezuela. The colloidal coal was obtained bymeans of the process described in section 2.1, then dried in an oven at 85 °C. The samples masswas varied from 180 to 750 grams of dry coal, either colloidal or conventional. Ignition was aidedby the addition of a few drops of kerosene. Flame time was measured from its onset to totalextinction.

A petroleum coke obtained from a refinery delayed coking process was also used in these tests inorder to further assess the combustion behavior of CCW. To this end, a 56.7% coke contentcolloidal suspension was prepared using the same process described in section 2.1. A drop of thesuspension was placed on a ring of 1 cm diameter; the suspension drop was small enough to be heldby surface tension and viscous forces on the ring without falling out. Then, a propane torch wasplaced underneath the ring containing the sample and the process of ignition and burning wasfilmed and evaluated.

2.2.2 Reburning studies. The reburning tests were conducted in GE's boiler simulation furnace(BSF) test unit, shown schematically in Fig. 2. The BSF has a firing rate range of 200 000 to 1 000000 BTU h−1 and it has been designed to simulate a coal-fired boiler. It consists of a burner,vertically down-fired radiant furnace, and horizontal convective pass. A variable-swirl diffusionburner with an axial fuel injector is used to simulate the temperature and gas composition of acommercial burner in a full-scale boiler.


6 of 25 4/9/10 3:17 PM

Fig. 2 Boiler simulation furnace (BSF) used forreburning tests of conventional slurry andcolloidal suspensions of Eastern bituminouscoal. Test facility of GEER, Santa Ana,California.

Numerous ports located along the axis of the facility allow access for supplementary equipmentsuch as reburning injectors, over fire air injectors, and sampling probes. The BSF is equipped withan electrostatic precipitator to control particulate matter emissions in a manner representative of afull-scale system and can be configured to match the residence time-temperature profile and furnaceexit gas temperature of a typical utility boiler. Facility instrumentation includes a system ofcontinuous emissions monitoring to sample O2, CO, CO2, NOx, and SO2. For reburning tests, NOxemission was the primary performance indicator. The combustion system was operated in such amanner as to achieve the target initial NOx concentration. The BSF operational procedure was tofire the unit on natural gas around-the-clock to maintain thermal conditions. At the start of each testday, a check was performed to ensure that the unit was operating at steady state, as evidenced bysteady emissions readings. A set of baseline measurements was taken prior to the first test run. Testruns were conducted, and then a repeat set of baseline measurements were taken to ensure thatconditions had remained constant throughout the tests.

The auxiliary systems used for the reburning tests were a slurry storage tank, continuous stirrer,peristaltic pump, slurry transport line, and twin-fluid atomizing nozzle. According to GE sexperience, high shear forces can cause slurries to separate during pumping, causing pooratomization and/or plugging of feed lines. Thus, a peristaltic pump was used to meter the coal waterslurry and the CCW suspension, because this type of pump generates lower shear force than otherdevices such as gear pumps. Further, long transport lines can cause slurries to separate, particularlywhen the lines are oriented upward. Therefore, the slurry (and CCW suspension) storage vessel waspositioned as close to the atomizer as possible. This approach helped to shorten the slurry transport


7 of 25 4/9/10 3:17 PM

line. The storage vessel was equipped with a continuous stirrer that was necessary to keep theconventional slurry in suspension.

An in-house manufactured twin-fluid nozzle was use to inject the sample in the reburn zone.Commercial nozzles tend to plug during coal slurry atomization. Compressed air joins with theinjected fuel through a tube in a Y-shaped configuration jet, atomizing the coal-water fuels innumerous small droplets at the injection point. The superficial velocity of the atomizing air at theinjection tip was estimated to be around 77 m s−1, a speed that would not cause the fuels to eitherimpinge upon the furnace walls or be propelled up into the flame zone, either of which wouldartificially impact on test results.

The atomization flow rate was held constant and the air-to-liquid mass ratio ranged fromapproximately 1.0 to 0.5 as reburn heat input varied from 10 to 20%. The ideal situation would beto hold both spray pattern and droplet size constant. However, this is not strictly possible when theliquid flow rate changes for different reburn rates. Functionally, droplet size typically decreases asatomization air flow rate (and air-to-liquid mass ratio) increases to a certain extent, and then levelsoff as additional air is added. The lower mass ratio evaluated, 0.5, is closer to the design point forthis type of nozzle while the higher mass ratio, equal to 1.0 for the lower reburn rates, generallywould not be expected to cause droplet size to decrease dramatically.

The method used was to run all points at the same atomization air flow rate, attempting tooperate near the flat part of this curve. This approach was intended to maintain the spray patternwhile minimizing the impact on droplet size. Specific atomization tests would be needed for anin-depth analysis.

2.2.3 Reburn fuel. Eastern bituminous coal was used to prepare CCW dispersions andconventional coal water slurry that were the reburn fuels for the test program. The conventionalgrind had a size distribution such that approximately 70% of the material passed through a US 200mesh sieve. Such a grind is typical for U.S. pulverized coal-fired boilers. The colloidal suspension,containing 60% Eastern bituminous coal, was prepared by Nano Dispersion Technology in Panamaand shipped to the GE facility at Santa Ana, California for reburn tests, where it was kept in storagefor several weeks before testing. The conventional coal slurry was prepared at GE Santa Ana,California facility, using the same Eastern bituminous coal, 200 mesh. For reasons described later,water content of the conventional slurry varied from 40 to 45%. Both samples, conventional andcolloidal, had the same type and amount of additives.

2.2.4 Test matrix. Two series of reburning tests were performed, including one with a colloidalcoal water slurry and one with a conventional coal water slurry. The coal dispersions, eitherconventional or colloidal suspension, were the reburn fuel, with natural gas as the main fuel. Afterstable emissions were verified during natural gas firing, the slurry was pumped and atomized intothe BSF furnace zone. NOx emissions were measured throughout the tests to determine theachievable NOx reduction. For selected test conditions, ash samples were collected from theconvective pass of the BSF and loss on ignition (LOI) was measured.

Test variables included: Reburning heat input (10 to 20%): varied by adjusting the reburn fuel flow rate. Reburn zone residence time (240 to 590 ms): varied by moving the over fire air injector

position. It is noted that varying the reburn zone residence time also varies the overfire air injectiontemperature.


8 of 25 4/9/10 3:17 PM

Initial NOx concentration (250 to 400 ppm): varied by adjusting burner conditions. Fuel water content (40 and 45%).

3.0 Discussion of experimental results

3.1 Particle size

After milling coal in the IKA® mill, particle sizes were, in general, evenly distributed from verylarge to very small sizes, i.e. mostly unimodal or slightly bimodal distributions. Fig. 3(a) is amicroscopic photograph of such a sample. According to the sieving tests, particle populationsbelow 20 m increased initially from 10–30% to 50–60% w/w after the wet-comminution process.Particles could be as small as 100 nm, depending on the type and content of coal, initial grind andprocess conditions such as viscosity prior to the wet-comminution, residence time and total energyinput. In general, it was found that both low viscosity and high coal content would improvecomminution and the best results were obtained for dispersions having coal contents between 55 to60%. In this range, coal content was sufficiently high and viscosity was still manageable.


9 of 25 4/9/10 3:17 PM

Fig. 3 Microphotograph of coal grinds, a)conventional grinding, b) wet-comminutionprocess. Particles below 1 m cannot be seenclearly in the picture (40× objective).

Microscopic observations confirmed the sieve results, while measurements with the Nanotracshowed a significant overall reduction in size of the colloidal population (<6.5 m). Fig. 3(b)depicts a microphotograph of the processed suspension. It can be seen that over a background ofvery small particles, there remains a few larger ones. Typical sieving results are shown in Fig. 4,which is a histogram of % weight retained between two sieves, for a 40 mesh ground sample (lessthan 400 m) and the same sample after the wet-comminution process. The population below 20 mincreased significantly as verified in the microscope. Few particles between 8 and 20 m wereobserved.


10 of 25 4/9/10 3:17 PM

Fig. 4 Percentage of mass retained between twosieves. Mesh size (in m) shown in the x axis isthe larger aperture sieve. Exception is datacorresponding to 20 m mesh size, the smallestsieve used; the histogram bars correspond topopulation below 20 m.

Additionally, Fig. 5 shows the results of the Nanotrac apparatus; the figure depicts the colloidaldistributions, before and after the wet-comminution process. The conventional grind produces acolloidal population that is below 30% weight and down to 0.7–1 m, while the wet-comminutionprocess increases the colloidal population up to 80% and decreases size close to the nanorange (100nm or 0.1 m).

Fig. 5 Particle size distribution, before and afterwet-comminution, of the colloidal sizepopulation.

3.2 Wet-comminution energy consumption

As mentioned in section 2.1, the power required for wet-comminution was measured and the results


11 of 25 4/9/10 3:17 PM

of these measurements are presented in Fig. 6 that includes values of a different process forcomparison purposes. The values reported take into account the power that would be necessary torun an industrial wet-comminution machine (i.e., net power, shaft and seal power); the particle sizeis the median of the distribution.

Fig. 6 Comparative energy consumption as afunction of mean particle size for coal wet ballmill and wet-comminution (Longlian et al.8).The legends indicate the concentration and typeof coal used for each process.

Wet-ball mill energy consumption for milling Illinois N° 6 coal (Longlian et al.8) are comparedwith wet-comminution of Eastern bituminous coal and Guasare coal (from eastern Venezuela). Ascan be seen in the picture, the wet-comminution process consumes the same order of energy toproduce a suspension with a very large population of colloidal particles that the ball mill uses forproducing say, particles of 100 m. This is an important result for any consideration ofmass-production of CCW at relatively low cost and in large quantities. This point will be addressedlater.

3.3 Colloidal coal suspension stability

Changes in CCW suspensions and conventional slurries were followed by laser backscatteringscans, as mentioned in section 2.1. Both types of fluid were placed in vials containing about 7 mLof each sample. Repeated scans were performed, beginning at the moment the samples were pouredinto the vials, and continued until no more changes were detected in the scan spectra.

The latter can be seen in Fig. 7(a) that also depicts both a photograph and a drawing of the vial(see legend for detailed explanation). The colloidal coal suspension was monitored until no morechanges were observed (214 h). Only a very thin layer of water at the top of the sample wasobserved. Further, no sediment layer was detected using the thin wire, after a few weeks of storage.


12 of 25 4/9/10 3:17 PM

Fig. 7 Percentage of light transmission along theglass vial for a) colloidal coal suspension and b)coal slurry. Solid content: 60%. In a), the dashedarrow shows the narrow region where light istransmitted through a thin layer of clarified,clean water. In b), the arrow points to a largerregion where water and some small coalparticles remain. Test conducted at roomtemperature.

The scans for the conventional slurry are shown in Fig. 7(b). This sample experienced a typicalsedimentation pattern, i.e., gradual formation of a layer of clarified water at the top that producedan increase in backscattering near the free surface (a ring of particles were trapped at the meniscus).The sample was monitored up to 121 h. After that, no more significant changes were detected. Afew weeks later, the thin wire test was performed and a thick layer of coal was detected at thebottom.

Samples of colloidal coal suspension and conventional slurries were stored in large glass jars forfollow-up. After months of storage, only a thin layer of water was observed at the top and nosediment was detected at the bottom. Conventional slurries had sedimented just a few hours afterthey were prepared. It is interesting to note that the colloidal coal suspension does not settle butexperiences some water separation at the top, in a manner similar to the syneresis of a gel. It is as ifthe whole structure shrank a little exuding some water.

Conventional coal slurries are very unstable with regards to particle settling and most typicalformulations must include stabilizers such as polymers or polymeric surfactants, as has beenreported by Boylu et al.,9 Dinçer et al.,10 Mishra et al.,11 Tiwari et al.12 and Ono,13 among otherauthors. These stabilizers behave as dispersants to reduce flocculation leading to settling, or asrheological behavior modifiers inducing viscoplasticity or a yield stress. According to Chen et al.14and Ono,15 another alternative to reduce settling is by increasing the colloidal particle population.Sedimentation is prevented because the colloidal fraction behaves as a pseudofluid with a


13 of 25 4/9/10 3:17 PM

viscoplastic behavior that suspends large particles inhibiting their settling. This is further discussedat the end of this section.

3.4 Rheological behavior and pipe handling of CCW dispersions

Fig. 8 depicts the apparent viscosity as a function of shear rate (0.01 to 100 s−1) for a 60% coalcontent colloidal suspension (Easter bituminous coal). It can be observed that viscosity decays fivefold and the high shear viscosity (100 s−1) reaches a value of 0.8 Pa s. The colloidal suspension alsoexhibits a viscoplastic behavior with a yield stress of about 20 Pa.

Fig. 8 Apparent viscosity as a function of shearrate for a colloidal coal suspension, 60% coal.Temperature: 25 °C.

The same kind of test was performed on conventional coal slurry (same coal type and content),but sedimentation occurred during the measurement and the resulting flow curves were deemedunreliable. However, the observed low shear viscosity was about 100 Pa s, the yield stress was lessthan 1 Pa and the high shear viscosity was about 0.5 Pa s.

The low shear viscosity of the CCW was high because particle interactions build up due to theincrease of the colloidal population. Forces between colloids induce the formation of large 3-Daggregates. However, high shear overcomes the interparticle forces and the size of the aggregates isreduced, thereby decreasing viscosity dramatically.

The colloidal coal suspension was certainly more viscous than the coal slurry. However, pipingthe suspension was an easier and more stable operation because no plugging or dynamicsedimentation occurred during the pipe handling of the conventional slurry. Pumping and tube flowof the conventional slurry and colloidal coal suspension with 55 and 60% coal content wasmonitored during the combustion tests described in sections 2.2 and 3.6. A Masterflex peristalticpump and a custom plastic hose were used to convey the colloidal suspension and conventionalslurry out of a 25 L container to the entry of the boiler nozzle, about 2 m downstream of the pumpdischarge. The pumping behavior was observed for several hours during combustion tests. Theconventional slurry, unlike the colloidal suspension, experienced frequent plugging that was onlyeased by increasing water content.

Further, the experimental evidence collected in a gravity driven capillary viscometer points tothe fact that the colloidal coal suspension slips at the wall, once the yield stress has been overcome.

Fig. 9 shows shear stress as a function of apparent shear rate for a set of data (each point is the


14 of 25 4/9/10 3:17 PM

mean value of several tests) collected for a 68.5% polymodal suspension of Guasare coal (largecolloidal population plus 1/3 of particles > 150 m). Viscosity calculated from the data is about 530mPa s, with the exception of the data collected with the 5 mm glass capillary. In contrast,conventional slurries or suspensions of the same coal content, but with a narrower particle sizedistribution, were very viscous and would exhibit very limited fluidity. Further, the polymodalsuspension did not settle, despite the presence of the larger particles.

Fig. 9 Shear stress as a function of shear rate fora polymodal suspension evaluated in a gravitydriven capillary viscometer at room temperature.Error bars: 10% for shear rate; 5% for shearstress. The slope of the dashed line is 530 mPa sThe Figure also shows a schematic drawing ofthe measuring device.

As already mentioned, the data collected with the 5 mm glass capillary did not follow the trendexhibited by the remaining data and the reason is discussed next. The behavior of the samplesduring the tests is worth recalling and discussing. Once the capillary outlet was unblocked, thesamples tested in large diameter capillaries would flow out almost immediately. For the smallercapillary no flow was observed in the first few seconds after unblocking the capillary exit, then flowwould start abruptly and be fast. This behavior was interpreted as follows (see eqn (1)). Small daccounts for low shear stress that can be below the critical stress to flow, or yield stress. This is whythe polymodal suspension does not flow under this condition. However, the ensuing behavior is nottypical of a yield stress fluid that eventually flows, albeit very slowly. Our suspension flowssuddenly and very fast. This can mean that the coal suspension lubricates at the wall and the delayto flow is the time required to migration of the continuous fluid to the wall. Once the lubricatedlayer reaches a critical thickness, the suspensions flows freely.

3.5 The notion of a pseudo-fluid

The behavior depicted in the previous section can be understood if the CCW is conceived in thefashion described next. Probstein16 advanced a model for the viscosity of concentrated bimodal


15 of 25 4/9/10 3:17 PM

suspensions of colloidal and non-colloidal particles. The important feature of this model is that thecolloidal fraction is assumed to act independently of the coarse fraction and to impart to a stablesuspension most of its non-Newtonian characteristics, such as shear thinning behavior. Largeparticles are essentially unaware of the existence of the colloidal particles. Instead, they see a stiffened single-phase fluid with the same rheological behavior as the colloidal suspension.

If we recall the particle size distribution presented in Fig. 4 for the colloidal suspension of coalin water, even though the system is not bimodal, there exists a large colloidal fraction and aspectrum of sizes for non-colloidal particles. In essence, it could be said this suspension iscomprised of a large colloidal fluid made of water, submicron and non-submicron particles. Thecomposite, water-colloidal particles act as a pseudo fluid that behaves as one liquid phase whichsuspends the non-colloidal particles and impedes settling.

3.6 Combustion of CCW dispersions

The evaluation of the colloidal suspension of coal in water as a fuel posed questions relating to thescale of the test. Testing the suspension as a boiler fuel in a pilot plant was ruled out because theeffect of particle size could be masked by the unrealistically uniform high temperatures thatdevelop at this level. In view of this, we decided to undertake some laboratory tests and a pilot testin reburning mode, so that the effect of particle size could be truly assessed. An additional dividendfrom the pilot evaluation would be the testing of the suspension handling in pump and accessories,as described in section 3.4.

From the standpoint of combustion, one of the main features of the colloidal suspension of coalin water is the great increase in surface area relative to a slurry made with a regular grind. Thelargest surface area would lead to a greater reactivity. To test this idea, flame lifetime of drycolloidal and regular grind coal was measured, as described in section 2.2.1. Fig. 10 shows theflame time in seconds as a function of mass of coal. Flame lifetime was almost double for thecolloidal coal in approximately the whole range of dry coal mass evaluated.

Fig. 10 Flame time of life as a function of coalmass for dry conventional coal grind and drycolloidal coal.


16 of 25 4/9/10 3:17 PM

The ignition of a wet 57.5% colloidal coke suspension was observed as explained in section2.2.1. We consider that the conditions in which this test was carried out were less favorable than theones found in a commercial boiler. In addition, petroleum coke has less volatile matter than coal.We found that the suspension would ignite and combust fast and completely. A photograph of thecombustion process (and experimental set-up) is shown in Fig. 11.

Fig. 11 Combustion of a hanging drop ofcolloidal coke suspension containing 42.5%water. The sample is ignited by means of apropane torch.

Regarding the pilot scale, it was decided that using the CCW as a reburn fuel was a good optionto assess the benefits associated with the colloidal suspension new attributes, such as particle size,pseudo fluid characteristics and superior handling. Reburning is a technology in which a fraction ofthe input fuel to a boiler is moved to a downstream location, establishing a fuel-rich reburning zonein which NOx emissions are reduced. To this end, we contracted GE-Energy to perform pilot scalereburning tests to define the NOx reduction and carbon burnout achieved with the suspension intheir Santa Ana facility. In GE's experience, the best way to run such tests is to comparativelyevaluate a baseline fuel and a test fuel, thereby highlighting the potential benefits of the test fuel.Therefore, for the subject work GE performed reburning tests with both a conventional coal grindand the colloidal coal suspension of the same coal type. NOx reduction and carbon burnout weremeasured as a function of key reburning test parameters.

3.6.1 Qualitative assessment. Based on the observation during reburn tests, GE made qualitativeassessments of both colloidal and conventional coal water slurries. The colloidal slurry wasprepared in Panama, shipped to Santa Ana, California, and was kept in storage for several weeks.Even after shipment and storage, it was found that the slurry did not settle in the containers andmaintained good flow conditions. While detailed atomization studies were beyond the scope of theproject, the colloidal suspension characteristics such as handling, pumping and atomization were


17 of 25 4/9/10 3:17 PM

good.The conventional slurry was prepared at the GE Santa Ana facility with 40% water by weight.

However, this slurry settled in the bottom of the storage container within a few hours. While GEdid not perform a detailed quantitative analysis of slurry atomization, poor atomization quality andfrequent plugging in the transport line was experienced with the conventional slurry containing40% water. Even after making several modifications to the transport lines and flushing the reburnfuel injector and transport line, handling and atomization quality were much poorer than for thecolloidal slurry. To provide a more direct basis of comparison with the colloidal slurry, theconventional slurry was then prepared with 45% water by weight and most tests were performedwith slurry containing 45% water. A few points with conventional slurry containing 40% water andcolloidal slurry containing 45% water were completed and provided a reference basis forcomparing results.

3.6.2 Comparison with other reburn fuels. Fig. 12 shows the performance of CCW suspension(40% water content) and conventional slurry (40 and 45% water) regarding NOx reduction as afunction of reburning heat input. On the one hand, the conventional slurry that had 40% watershows the worst performance, and this was likely caused by the handling and atomization problemsmentioned before. On the other hand, it can be said that the CCW suspension exhibited the bestperformance considering that water content was less (40%) than the conventional slurry (45%) thatshowed similar behavior at low heat input (recall heat de-rating due to water). The latter figureplaces the current CCW suspension in the context of two other reburning fuels that have been testedat the BSF, natural gas and dry bituminous coal. As expected, natural gas is the best fuel in terms ofNOx reduction, while dry coal is as low as the conventional slurry with 40% water. It can be saidthat:

Fig. 12 Comparison of reburn performance forvarious fuels, including conventional coalslurries having 40 and 45% water, and colloidalcoal suspensions obtained by means of thewet-comminution process.


18 of 25 4/9/10 3:17 PM

Natural gas is the most reactive of these fuels due to its ability to readily disperse and react.However, natural gas is also typically the most expensive reburning fuel, and thus there iscommercial interest in utilizing other fuels such as coal for reburning.

The dry bituminous coal has relatively low volatile matter and high fixed carbon, and is nothighly reactive.

The colloidal suspension was generated using a bituminous coal that, on its own, would not beexpected to be highly reactive. Its performance in reburning is nevertheless only surpassed bynatural gas. This has to do with the increased reactivity of the coal particles owing to their size.

Loss on ignition (LOI) was also evaluated for the CCW suspensions (45 and 40% water) and theconventional slurry (45% water). LOI was measured for two residence time (240 and 590 ms) andtwo levels of NOx input (250 and 400 ppm). There was no significant difference between the twotypes of dispersions and LOI varied from 3.2 (590 ms, 400 ppm NOx) to 1% (250 ppm, 240 ms),depending on NOx level and residence time.

4.0 Discussion

In what follows we advance some relevant questions concerning the massive use of colloidal coalsuspensions in water for power generation. We have separated them into three types of questions,mass production, potential applications and uses of the material in the future, and researchquestions. These have to do with particular issues that need to be addressed in the short term, whichwould open windows of opportunity for tangible clean coal applications.

4.1 Mass production feasibility estimation

Any commercial application involving the use of CCW would require the manufacture of largevolumes of the product. When scaling up a regular mixer, shear levels and specific energy inputsare hard to reproduce at larger scales from, say, lab conditions. This does not happen with thewet-comminution technology described here.

As mentioned earlier, the wet-comminution process takes place in a hydrodynamically inducedfluid region that traps the particles and subjects them to high shear. This region is confined to afilm and the film can be reproduced at large scales with the same dynamic conditions (it should benoted that the ratio of film thickness to vessel diameter is equal in all scales). This is what ensuresscale up success. Another interesting feature of this technology is that residence times are small.The latter tend to vary with the material being processed but are seldom greater than 10 s. Again,that stems from hydrodynamic focusing, which traps particles and mills them by contact forces in ashort time. The combination of short residence time with hydrodynamic focusing allows for an easychange of scale and facilitates large production of CCW.

The ultimate successful application of colloidal coal in water suspensions would rely on largesuspension output capacity at reasonable energy expenditure and additive costs. To assess the latterissue, we have evaluated the general steps associated with the mass production of the material.These are presented schematically in Fig. 13. In terms of specific energy required formass-producing CCW, three stages are involved. First, raw coal must be milled to less than 0.4 mmparticle size (40 mesh), and combined with water and additives to prepare the pre-slurry fed to thewet-comminution device. Commercial energy usage values to achieve this target particle size areabout 10–12 kW h ton−1, according to on-line publications of Xstrata Corporation


19 of 25 4/9/10 3:17 PM

(www.xstrata.com) and Khodakov.17 The pre-slurry formation is carried out by regular mixingequipment and typically consumes less than 4 kW h ton−1. The wet-comminution process consumesabout 120 kW h ton−1 of slurry leading to colloidal coal. If these factors are added, a figure of lessthan 150 kW h ton−1 is obtained, which is the same energy level most ball mills use to grind coal toabout 80 to 100 m.

Fig. 13 General process for manufacturingCCW.


20 of 25 4/9/10 3:17 PM

Another point should be made here regarding colloidal coal. Advanced technology grinders,such as Xstrata Isamill can manufacture commercially micronized coal with great efficiency.However, this type of mill faces difficulties going to submicron size because these fines (aspeople call very small particles in the mining industry) float in the air and become a health hazard.Colloidal coal particles are better handled in the presence of water.

The second aspect for mass-producing CCW has to do with processing capacity. A commercialwet-comminution device has, typically, 16 L of vessel capacity and operates at residence times ofless than five seconds. When scale-up factors are considered, a manufacturing plant can bedesigned for a capacity of 500 000 ton/year of CCW with less than five units. These devices are notvery large in size and can be easily accommodated in any industrial facility.

The third element associated with mass-producing CCW is very important and has to do withoperational costs that include cost of raw coal ( 58 US$/ton), additives (10 US$/ton) andmanufacturing (7 cents/kW h) costs. The estimate for the production of a suspension containing60% of coal is about 3.2 US$/MBTU produced (or 90 US$/ton of CCW). The last figure takes intoconsideration the heat de-rating due to water evaporation (or LHV). The latter is reasonableconsidering that additional advantages such as ease of handling (no settling, no trouble pumpingand atomizing, good flow rate control) and better reactivity are not factored in the latter figure.

4.2 Potential uses of CCW

As pointed out by Ansolabehere et al.,1 CO2 capture and sequestration is the critical enablingtechnology that would reduce CO2 emissions significantly while also allowing coal to meet theworld's pressing energy needs .

Within this context, ultra supercritical boilers, various combined cycle gasification schemes andoxy-combustion are the most important and promising technologies being evaluated and usedtoday. While these initiatives are fine, a long-term vision for coal might require a paradigm shift,like producing electricity in distributed configurations with greater thermal efficiency as a standard.What if coal is utilized with virtually no CO2 emitted? Firing coal in more efficient devices such aslow speed diesel engines, gas turbines and even fuel cells could greatly reduce CO2 emissions.

As already mentioned, all of these applications require various degrees of coal cleaning. Now,according to Cooper,18 coal cleaning is not a new activity but treating coal to the extent it can beused in a direct carbon conversion fuel cell (DCC) is. The same applies, albeit to a lesser extent, togas turbines and diesel engines.

Deep coal treatment for DCC involves the reduction of ash and pyrite by mechanical andchemical means to less than 1%. Similar targets are required for gas turbines and to a certain extent,to diesel engines. All coal-cleaning strategies are affected by particle size. Surface treatments arestrongly impacted by surface area and flotation forces depend directly on particle size. Colloidalcoal could represent a big step in improving deep coal. It could facilitate the removal of mineralmatter within the wet-comminution process and help in surface treatments. How this product can beincorporated into coal cleaning procedures is an important question in the quest for enabling the useof coal in the future.

Coal cleaning is a required step but not the only one. Efficient energy conversion is another,which is associated with something that can be called fuel preparation. Colloidal coal can beviewed as a form of fuel preparation. Fuels need to be handled and must be the right fit for the


21 of 25 4/9/10 3:17 PM

conversion device being used. Sekar et al.19 fired low-grade diesel fuel in a diesel engine withoxygen enrichment and water. Their idea was to reduce NOx and particulates by firing with watermodulating the high temperatures associated with oxygen enrichment. Similarly, colloidal coalcould be used in a low-speed diesel engine in the presence of oxygen with great efficiency. Fuelinjection control would be greatly improved relative to any slurry by the homogeneous nature andrheological behavior of CCW. The same situation could apply to a gas turbine. For the case ofDCC, the match with CCW is yet to be determined but could yield great benefits.

We have made arguments concerning CCW in low-speed diesel engines, gas turbines and evenfuel cells, but colloidal coal could also benefit gasification schemes and oxy-combustion.Integration of gasification into the total IGCC plant imposes additional considerations on thetechnology (Holt et al.20). Moving bed gasification, for example, cannot deal with a significantfraction of coal fines, which means that 20–30% of the processed coal cannot be fed into it (notethat CCW can also be manufactured with low-rank coals which are commonly used in these type ofunits).

A question then arises as to the potential impact of CCW in this technology. High temperature,entrained-flow gasifiers do not have these issues and are thus more readily integrated into an IGCCsystem. High-pressure operations are favored in these units. The introduction of coal into apressurized gasifier can be done either as dry coal feed through lock hoppers or by slurrying thefinely ground coal with water and spraying it into the gasifier. The latter introduces 30% w/w liquidwater, which is desirable if the coal has low moisture content. For high moisture content the gasifierfeed can approach 50% water, which increases the oxygen required to gasify the coal and vaporizethe water and reduces the overall operating efficiency. Colloidal coal could enter into thistechnology in two ways. For low moisture coal it would facilitate handling and flow control into thegasifier by virtue of its non-settling attribute. In the case of high-moisture coal, the very process ofwet-comminution could contact and extract some of the water present in the coal and maintainwater levels close to 30%, avoiding the efficiency de-rating.

In oxy-combustion schemes, CCW could have an impact as a result of the increased reactivityassociated with small particles. That, in turn, could change the amount of oxygen required foroptimum operation with re-circulated flue gas.

4.3 Research questions

The use of colloidal coal poses many specific questions that require research efforts. The main setof questions involves making CCW usable in diesel engines and gas turbines. Table 1, afterWibberley et al.,2 lists current specifications for using coal water fuels in these machines.

Table 1 Current specifications for coal water fuels according to the application (Wibberley et al.2)

Property Current CWF for boilers Gas turbine Diesel engine

Mean particle size/ m 10-20 4–6 5–15

Particle size, m % mass passing

−5 m 5 60 30


22 of 25 4/9/10 3:17 PM

Particle size, m % mass passing

−10 m 35 100 60−20 m 50 100−75 m 75−250 m 95−500 m 100

Coal content (% w/w) 65–70 55–60 50–55

Viscosity/mPa s @ 100 s−1 500–1000 400 300

The colloidal coal suspension advanced here is not far from meeting these specifications withinreasonable cost. However, many questions emerge from this table and the foregoing questions mustbe addressed in the near future.

For colloidal coal, would the particle size shown in the table be really required? Wouldn't the colloidal particles greatly enhance the reactivity of larger particles? What are the detailed combustion characteristics of CCW? Would coal in colloidal form require deep cleaning? What would be the ideal water content for a colloidal coal suspension to be used in gas

turbines and diesel engines?Research questions also emerge from gasification and oxy-combustion applications. Based on

the previous discussion the following are specific areas that need be assessed, Can CCW be applied in a moving-bed gasifier? Can CCW be used to manufacture fuels and chemicals through gasification? Would the wet-comminution process contact and extract water present in high-moisture coal? What is the impact of overall operating efficiency in a gasifier as a result of improved flow

control using CCW? Can CCW optimize the water content into the gasifier without increasing the viscosity of the

suspension? Can CCW help reduce the amount of oxygen required in an oxy-combustion scheme?

Given the prominence given by many industrial nations to IGCC in order to reduce greenhousegases, the foregoing questions are of great importance and are also potential good opportunities toimprove these processes.

5. Concluding remarks

All evidence collected so far points to the surfacing of a new material for power generation. Thismaterial is not a slurry, as slurries have been known so far. Power generation is a delicate matter forthe world. Many drivers are involved and some of them point in opposite directions. Of all drivers,two can be considered invariant in the foreseeable future. One is the continuous use of coal as fuelfor power generation. As mentioned in the introduction, this has to do with coal being cheap andabundant. The other driver is global warming.


23 of 25 4/9/10 3:17 PM

We have advanced a colloidal coal in water suspension that has outstanding features such aslong-term stability to settling and very large surface area. We have tested the combustion of thismaterial in reburn mode with excellent results. Reburning is a very harsh combustion application,which typically takes place under sub-stochoimetric conditions and low flame temperature. Thismaterial has various potential applications that could open new doors for using coal in anon-pollutant manner. This work also poses research and vision questions that if answered, couldimpact greatly the way we produce electricity today.

List of acronymsBSF Boiler simulation furnaceCCGT Combined cycle gas turbines

CCS CO2 capture and sequestration

CCW Colloidal coal in water suspensionsDCC Direct carbon conversion fuel cellIGCC Integrated coal gasification combined cycleLOI Loss on ignitionLSD Low speed enginesUSC Ultra super critical pulverized combustion

Acknowledgements

We are grateful for the invaluable help of Alonso Alvarado, Edgar Marin and Francisco Chang whocarried out most of the experimental work. We are also grateful to Professors L. S. Fan and R.Probstein for reading an earlier draft of this paper with a critical eye. The applied math division ofthe NSF, under ARRA, supported the work of D. D. Joseph, in part.

References

1 S. Ansolabehere, J. Beer, J. Deucht, D. Ellerman, J. Katzer, J. Friedman, H. Herzog, H. Jacoby,P. Joskow, G. Mcra, R. Lester, E. Moniz and E. Steinfeld, in The Future of Coal: Options for aCarbon-Constrained World, Massachusetts Institute of Technology, 2007, ISBN978-0-615-14092-6.

2 L. Wibberley, D. Palfreyman and P. Scaife, Efficient use of coal water fuels, CSIRO EnergyTechnology, Australian Government's Cooperative Research Centres Program, Pullenvale,2008, Report 74.

3 K. Thambimuthu. , Developments in coal-liquid mixtures, Center for Applied EnergyResearch, University of Kentucky, Report 69, 1994.

4 J. Simeon, Economical potential of coal-water mixtures, ICEASIE 8, IEA Coal Research,London, 1985.


24 of 25 4/9/10 3:17 PM

5 K. A. Davis, J. R. Valentine, E. G. Eddings and M. P. Hea, Evaluating the effects of low-NOxretrofits on carbon in ash levels, Proceedings of the Reaction Engineering International MegaSymposium: EPRI-DOE-EPA, Combined Utility Air Pollutant Control Symposium, Atlanta,1999.

6 US Department of Energy, National Energy Technology Laboratory, Micronized coalreburning demonstration for NOx control: a DOE assessment, Report DOE/NETL 2001/1148,2001.

7 F. Boylu, G. Atesok and H. Dincer, Fuel Process. Technol., 2004, 85, 241–250 [Links].8 C. Longlian, A. Liqian, W. Gong and J. A. Hejin, Fuel, 2007, 86(5–6), 750–757.9 F. Boylu, G. Atesok and H. Dincer, Fuel, 2005, 84, 315–319 [Links].

10 H. Dinçer, F. Boylu, A. Sirkeci and G. Atesok, Int. J. Miner. Process., 2003, 70, 41–51 [Links].11 S. K. Mishra, S. B. Kanungo and Rajeev, J. Colloid Interface Sci., 2003, 267, 42–48 [Links].12 K. Tiwari, S. Basu, K. Bit, S. Banerjee and K. Mishra, Fuel Process. Technol., 2004, 85, 31–42

[Links].13 2000.14 J. Cheng, J. Zhou, Y. Li, J. Liu and K. Cen, Fuel, 2008, 87, 2620–2627 [Links].15 T. Ono, Kona, 2003, 212, 202–207.16 R. F. Probstein, in Physicochemical Thermodynamics, Wiley Interscience, New York, 2nd

edition, 1994, chapter 9, pp. 298.17 G. S. Khodakov, Thermal Engineering, 2007, 54(1), 36–47 [Links].18 J. Cooper, Direct conversion of coal and coal-derived coal cells in fuel cell, presented at The

Washington Coal Club, Lawrence Livermore National Laboratory, Livermore CA, 2004.19 R. R. Sekar, W. W. Marr, J. E. Schaus, R. L. Cole and T. J. Marciniak, Proceedings of the 25th

Intersociety Energy Conversion Engineering Conference, 1990, 4(12–17), 320–325.20 N. Holt, G. Booras and D. Todd, Proceedings of MIT Sequestration Forum IV, Cambridge,

Mass., 2003.

This journal is © The Royal Society of Chemistry 2010


25 of 25 4/9/10 3:17 PM

colloidal coal in water suspensions

Documents