BNL-23860

THE USE OF ¡4ANGANESE SUBSTTTUTED

FERROTTTANTUM ALLOYS FOR ENERGY STORA,GE*

by

J. R. Johnson and ,J. J. ReillyBrookhaven NatÍona1 Laboratory

Department of Energy and EnvironmentDivision of Molecular Sciences

Upton, New York L7973u"s.A.

For publication in:Proceedings of the International Conference

on Alternative Energy SourcesMiami Beach, Florid.aDecember 5-7, L977

*üIork performed under the auspices of the u.s. Department ofEnerg,y.

BNL-23860

THE USE OF ¡4ANGANESE SUBSTITT}TED FERROTITANTUMALLOYS FOR ENERGY STOR.AGE*

J. R. Johnson and J. ,J. Rei1lyBrookhaven National Laboratory

Upton, New york, U.S.A.

ABSTRACT

Metal hydrides provide an attractive means of energy storage fora variety of applications. Titanium iron hydride, to date, hasreceived the most attention and shourn the most potential. Thephysio-chemical characteristics of this hydride can be atteredby the substitution of a third transition metal, such as manga-nese, yielding aIloy-hyd.rogen systems with a wide range of prop-erties. This allows hydride materials to be "tailor made" fora particular application.

Experimental results are presented and discussed in areas deal-ing with properties of major practical importance in the utili-zation of manganese subst,ituted ferrotitanium al1oys as hydrogenstoragie media. These areas encompass the following: (1) pres-sure-composition-temperature characteristics, (2, particle attri-tion properties, (3) effects of long-term cycring on alloy sta-bility, (4) ease of aetivation and reactivation, and (5) effectsof contaminants on al1oy activity. rn addition, the advantagesand disadvantages regarding the use of ternary alloys vis-a-vistitanium iron will be discussed, as will the development of anoptimum ternary alloy for use with a particular peak shaving op-tion, i.e., the regenerative id,Z-C.l-Z system.

II TRODUCTION

Many metal alloys react dírectly and reversibly with hydrogengas at ambient temperatures and reasonable pressures to formsolid hydride compounds [I-9]. Certain of these materials havebecome a focus of attention in the past few years because oftheir potential as energ]¡ storage media. A multitude of appli-cations have been suggested for such compounds, both mobile and

*l{ork performed under the auspices of the u.s. Department ofEnergy.

-1- JRJ,/JJR

stationary in nature. Most of the proposed applications of thesea1loy-hydrogen systems derive their attract,iveness from compari-son to the conventional storage of hydrogien as a compressed gasor a cryogenic liquid and in comparison to competing energy stor-age options such as batteries f1O-12]. Ultimately the advantagesor disadvantages of the metar hydride energv storage compoundsare d.irectly rerated to their physío-chemical properties.

To date' the metal hydride which has received the most attentionand sholvn the most potential for practical application has beentitanium iron hydride. This has resulted principatly from itsfavorable pressure-composition characteristics, production costsand the relative abundance of the starting materials. This met-al-hydrogen system was discovered and characterized at Brook-haven National Laboratory (BNL) tr] and an engineering scalereservoir containing this material was constructed at BÀtL foruse by Public service Erectric and Gas company (psE&G) of NewJersey. It was combined with a water electrolyzex and fuel cel1unit as a demonstration of a road leveling application [r¡].The unit has been operated successfully over 6O charge-dischargecycles and is currently on loan to BNL. vtith respect to mobileapplications, several organizations have converted a variety ofvehicle types for operation with hydrogen derived from titaniumiron hydride beds If¿-fA].

As has been noted previously [fZ], titanium iron hydride is notthe ideal storage material for use in all applications and theability to modify its properties is desirable. changes in theproperties of titanium iron hydride¡ âs well as certain otherhydrides, may be affected by the addi-tion of small amounts ofother components [fg], by variation ín the atom ratio of titan-ium and iron in the starting alloy [1,12] or by the partial sub-stitution or addition of another transition metal for iron re-sulting in a ternary a1loy of general composition TiFer,.IvI", l_fZ].The effects of variation in the starting iron,/titanium--råtio andpartial substitution of a third transition metal for iron areshown in Figs. 1 and 2. The partiar substitution of iron by an-other transition metal makes it possible to "tairor make", to ahigh deg:ree, ãrr aIIoy with appropríate properties for particu-lar energy storage and conversion applieations.

In a previous paper llZl. the integration of such ternary alloyswith energy storage systems, their relative advantagies, and, alsosome general properties, mainly pressure-composition character-istics, 'h¡ere discussed. rt is the purpose of this paper to

JR.T,/JJR -2-

discuss certain physio-chemical properties of a specific classof ternary TiFe*Mo alloys where M = manganese. The propertiesto be discussed aie all of practical importance in the utiliza-tion of these solid hydrides as energy storage media and are asfollows: (1) Pressure-composition characteristics includinghysteresi-s phenomena, (2) particle attrition as a function ofcycling (hydriding-dehydriding), (3) ease of initial activation(hydriding for the first time) and subsequent reactivaÈion aftercontamination, (4) effects of poisons on activity, and, (5) ef-fects of long-term cycling on the reversibirity of the absorp-tion-desorption reaction. Such properties were also of particu-lar importance to a current program at BNL coneerned with thedevelopment of an electrochemically regenerative hydrogen-chlo-rine system for energy storage. The development of an optimumhydride storage compound for use with this HCl system will alsobe discussed in relation to the overall reference design condi-tions for the unit.

The use of manganese as a substituent is particularly aËtractivesínce the metal is relatively inexpensive (-$0.50-0.6O/Lb') andabundant. cost increases on the order of 3-4 percent could beexpected for ternary alloys with 15 wt /o manganese. This isconsidered a small price to pay for the flexibility and increasein performance gained by such a substitution.

EXPERTMMTTAL

A1loy samples used in these studies r',rere prepared with startingmetals of at least 99.9 percent purity. All materials toreremerted ín an arc furnaee under an inert gas atmosphere exceptthose designated as rNco samples whj-ch were prepared at the rn-ternational ¡¡ickel company either by vacuum induction meltingin graphite crucibles or air melting in clay-graphite cruciblesfollowed by mischmetal* deoxidation [zo]. The INco samples wereprePared as large melts and should be representative of commer-cially available alloys. The compositions indicated for the al-loys are nominal values except for the INCO samples whose compo-sitions have been determined by elemental analyses. The nominalcompositions however should not deviate appreciably from theactual compositions sought since in the melting procedure extra

*Mischmetal is a low cost (S3.50,/1b) a1loy of various rare earthelements, comprising 4B-5O wt % ce, 32-34y" I¡a, I3-19l" Nd, 4-S%Pr, and -1.5% others.

-3- JRJ,/JJR

manganese (the most volatile component of the arloys) rn¡as addedto melts to compensate for any loss upon melting. None of thesealloys were homogenized or annealed unless specifically.indicated.

The laboratory apparatus and procedure for investigating pres-sure-composition properties of metal hydrides has been describedin detail previously [r,rg]. The same system was used in de-termining minimum initial activation conditions for all samples.The poisoning and particle attrition experiments vrere carriedout on a cycling system built primarily to investigate ehangesin the activity (absorption-desorption properties) of alloyà to-ward pure hydrogen as a function of the number of hydriding-de-hydriding cycles. This cycling apparatus has three experimentalstations attached to a pressure-vacuum manifold. A schematic ofone such station including a standard sample reactor is shown inFig. 3. The reactor for cycling experiments is essentially thestandard high pressure reactor used in pressure-composition ex-periments which had been jacketed with l-inch o.D. staintesssteel tubing to allow heating and cooling of the samples. Theínlet-outlet ports for the heat transfer fluids were connectedby 0.25-inch copper tubing to two sets of time operated solenoidvalves. One set of these valves operates in the cooling (hydrid-ing) mode whereas the other allows for heating (dehydriding).These solenoid valves are in turn connectgd to two_circulatingconstant temperature baths operated at -0o and 10ooc. Theerapsed time for each half cycle (heating or cooling) is approx-imately six minutes yielding a total cycle time of. 12 minutes.This system is operated continuously which results in a totalof 120 cycles per day. The remainder of the apparatus consistsof a reservoir of calibrated volume (-500 ml) attached by a0.25-inch stainless steel rine arso of known volume. The pres-sure is recorded using a O-IOOO psia Dlmisco pressure transducerconnected to a dual pen Honelnuell Electronik 196 chart recorder.The temperature is measured by a thermocouple in the bottom ofthe reactor and is recorded as a second trace on the same re-corder that monitors pressure. A sample or the entire stationcan be isolated from the manifold by means of bellows-seal valves.

Minimum initial activation experiments were performed to ascer-tain the ease of activation of the arloys. Approximately 5-10grams of an alloy was crushed and passed through a 2O-mesh(841 r¡m) stand.ard screen then loaded into a high pressure reac-tor. The reactor was conneeted Èo the high pressure-vacuum man-ifold and evacuated to about 1o-4 ïìm pressure for a period oftime usually on the order of 0.5 hr either with or without

JRJ,/J;TR -4-

heatíng. rn the former case the sample temperature was main-tained with an electrical resistance furnace or constant temper-ature bath. rn most instances the reactor was not heated orcooled but simply outgassed at room temperature. Samples Ìrerethen exposed to hydrogen usually at 500 psia (3A atm) . Thepressure was recorded as a function of time. The free gas vo1-ume and sample size in these experiments were such that changesof pressure not in excess of 30 psia were expected even if thealloys absorbed to their highest capaeity. upon completion ofan experiment the samples hrere isolated from the remainder ofthe manifold, cooled to liquid nitrogen temperature, and evacu-ated. The hydrogen content was then determined by heating thesample to 450"c and measuring the amount of evorved gas. Re-sidual hydrogen, that which did not evolve at 450oc, was mea-sured on a knor,rn amount of al1oy by a thermal decompositiontechnique whereby the sampre is heated to approximately lOoOoc.

Physical deterioration experiments (loss of sample activity withcycling) !ùere accomplished by crushing the a1Ioy samples and re-taining a 5-10 gram portion which had been sieved through a 2a-mesh (841 ¡lm) screen. 'The material was then activated in thenormal manner tf] and charged to a high H,/M ratio. ït was sub-sequently connected to a position on the cycling apparatus. Thefree gas volume above the bellows-seal valve to the reactor wasevacuated and hydrogen r,rras admitted. The hydrogen used was goldlabel quality Matheson gas of a minimum gg.ggg percent purity.The same hydrogen gas rnras used in each cyc1e. once the gas vo1-ume had been pressurized the sample position was closed. åo theremainder of the systemo the reactor valve opened and cycling wasbegin. The pressure swing on cycling could be adjusted, withincertain limits, by varying the system pressure, the cycle time,and the change in temperature. The change in composition percycle can be carcurated by noting the maximum and minimum pres-sures (^P). Any essential changes in the properties of the sam-ples, e.9., phase separatíon, were indicated by a change'in (^p)per cycle.

Particle attrition (particle size reduction) investigatíons weredone ín the same manner with the exception that a 5-10 gram sam-ple of a1loy was screened to a size range af -!o/+16 (1200 to20O0 pm) and then rescreened after successively increasing num-bers of cycles. Particle sizes smaller than 325 mesh (aa p,m)were not analyzed further.

-5- ,JRJ,/JJR

Poisoning experiments were performed using the same technique-The activated sample was isolated on the cycling apparatus anda mixture of hydrogen containing a specified amount of a con-taminant was admitted into the large free gas volume above thereaetor. The reactor valve was then opened and the cycling op-eration begun. Any change in activity could be readiry notedby a change in (^P).

RESULTS AND DISCUSSION

Pre s s ur e-compos i t ion-Tempe ra turç cha ra cter i s t ic s

In a previous paper lfZ] tåe effects of substitution or additionof another transition metal on the pressure-composition ísothermsof TiFe were presented. All such composition adjustments led toa metal-hydride system of increased thermodlmamic stability (iso-therm-pressures vtere shifted to lower pressures at comparabLe ll/Mvalues and temperatures) and generally to the introduction of adistortion or slopíng effect in the isotherms. Figs. 4-7 illus-trate this occurrence for TiFe systems subsÈituted with cobalt,chromium, copper, and molybdenum. The substitution of manganeseis no exception to these general observations as has been illus-trated in Fig o 2. It should be noted that as the content of ¡4t1

increases hydride stabilities increase and up to a compositioncorresponding to T_i_FeO.gOh6.2g Í-tt wt.% ¡f¡:) no loss of hydrogencapacity occurs. Ho\n/ever, upon increasing the manganese concen-tration of the alloy to fifeg.70¡{n0.30 a slight loss of hydrogencapacity is evident" This undoubted.ly results from a significantincrease in the amount of a second phase, Ti(Fe,l{n)2, present inthe "as cast" aIloy. Partial ternary phase diagrams for theTi-Fe-Mn system for "as cast" and equilibrium composítions havebeen developed by Sandrock [ZO] and Murakarni and enjyo lZtl, andthey should be consulted by the reader if further information isrequired.

The increase in stability of the manganese substituted ferroti-tanium alloy-hydrides, ërs with other substituted hydrides,may be explained qualitatively in terms of a correlation devel-oped by Lundin et aI. \ZZ,Z3] and the ',Rule of Reversed Stabil-ity" formulated by van Mal et at. [g]. The former correlationstates that a linear relationship exists between the standardfree energy of formation of a hydride (or logarithm of the p1a-teau pressure) and the interstitial hole size of the metalliccompound lattice. lFtre larger the hole size or the atomic volume,the more stable will be the resulting hydride. The correlation

JRJ,/JJR -6-

for each structure type is different and must be establishedfor each separately. The "Rure of Reversed stability" statesthat the more stable an intermetallic compound, the less stablewill be its hydride. A1r the manganese alloys ínvestigated doshow an increase in celr lattice parameter tão], âo, over thatfor TiFe and therefore may be expected to have tarler intersti-tial hole volumes resurting in *ãr" stable hydride phases.

As a result of the foregoing, it should be possible to producemanqanese substituted ferrotitanium hydrides with a wide rangeof pressure-composition properties which retain high hydrogenstorage capacity. This capability adds a degree oi rlexibilityto the TiFe system and is of practical imporiance since it per-mits one to "tailor make" a storage a11oy for a particurarapplication.

I{any of these manganese systems exhibit sloping prateaus. rngeneral this phenomena is composition dependent ánd results fromnon-equilibrium solidificat.ion effects. Such solidificationprocesses arise from large differences in liquidus-solidus tem-peratures in the phase equilibria and may give rise to segrega-tion within a single phase region (concentration gradients) ò,more generally to coring. This may be eliminated in many casesby homogenization-annealing of the starting intermetallic andsuch a case is shown for the TiFeg.zoho.l3-Hx system in Fig. g.The annealed sample was heat treaÈeå-in-å-iesïstance furnaceunder an inert atmosphere for three days at loOooc prior to hy-driding. Of course, Ëemperatures and. hold, times at temperaturemust be optimized individuarry for alroys of differingcompositions.

Finally, it has been noted that hysteresis is reduced in the man-ganese systems. An example of this is depicted in Fig. 9 in com-parison to TiFe. This effect eonstitutes an advantage for theuse of these materials in certain applications since lower pres-sures are required to charge a sample to an equivalent H/Nl va1ue.

Particle Attrition Studies

upon absorption and desorption of hydrogen the average particlesize of an alloy material decreases. This may lead to paekingof the hydride bed, in practical storage applicationsn largepressure drops in a flovv through system, and solid-gas separationproblems lZ+-zø7. TABT,E r illustrates the rate of particle sizereductj-on for TiFeo.zoho.tg (rNco #1) versus that for tire. The

-7- JRJ,/JJR

salient features to be noted here are that the rate of attri-tion of the manganese substituted ferrotitanium a11oy occursmuch faster than for TiFe and most of the attrition takes placein the first few cycles.

The rapid particle attrition rate of the "as cast" manganesealloys can be rationalized partially in terms of the microstruc-turar propertíes of these materials as has previously been sug-gested [ZO]. The "as cast" aIIoys generally are not singlephase materials and in some instances (the rNco sampres) havebeen deoxidized using mischmetal thus introducing an added phaseconsisting of rare-earth oxides. Ðepending on the compositionof the alloy the hydridable singre phase gi(re,¡tn) materiar canbe accompanied by hyd.ridable Ti solid solution, Ti(re,un)2 anda small amount of rare-earth metal oxides. The presence of Tisolid solution results in the formation of TiHz which facili-tates wedge cracking of the other phases whereas all other ex-traneous phases present provide pathways or nucleation siteswhich also greatly increase the rate at which a virgin al1oycan be activated.

Effects of Lonq-Term Cvcling on Allov Stabilitv

Titanium iron hydride and. analogous compounds are thermodlmami-cally metastable with respect to the formation of TiH2 and TiFe2.A series of long-term cycling experiments were therefõre con-ducted to investigate the long-term stability of tire and vari-ous manganese alIoys. The results of a representative portionof the experiments are given in TABLE rr. No change in the ac-tivity (deterioration) of any of these alloys was observed as afunction of time over a large number of cycles (20r000-30,000)as measured by the constancy of

^ (H,/Ti) . Further, in the sam-

ples of TiFe there was no evidence for the formation of TiH^or TiFe, in the alloys during the course of the experiment ásdetermined by x-ray diffraction analysis. The manganese al1oyswere not so analyzed, but since no deterioration in performancewas observed, it is highly unlikely that any disproportionationor phase separation is occurring.

The prime reason that no deterioration (phase instability) \^rasobserved for the hydrogen cycled fife and TiFe*t{n.r, alloys isprobably a direct consequence of a rather simple iule, relatingthe reactj-on of intermetallics with hydrogen, which is firmlybased on thermodlmamic and structural principles LZZ-2g7. Itmay be stated as follows:

JR/JJR -B-

If the metal atoms of an intermetallic compound are notmobile {as is the case in low temperature reactions} onlyhydride phases can result which are structurally similarto the starting intermetallic compound.

These cycl{ng experiments r,.rere carried out at lovr temperatures(-1 to 100"C) and therefore, although the al1oy phases are met-astable with respect to the formation of TiH2 and other phases,no disproportionation occurs because the metal atoms are notsufficiently mobile.

Ease of fnitial Activation Studies

The activation (first complete hydriding) of pure fiE'e ís timeconsuming, requires pressures on the order of 65 atm and temper-atures >300"C. These conditions require the use of materialsfor hydlide bed reservoirs which are expensive and involvestringent design constraints. Further, the alternative of acti-vating the a1loy in a separate vessel and subsequently transfer-ring it to the reservoir is not attracLive. Any simplificationsin the activation procedure which would moderate the conditionsnecessary and reduce the time reguired are important since sub-stantial reductions in costs could be obtained. Such simplifi-cations can be realized by the use of ternary manganese substi-tuted ferrotitanium alloys as all these materials studied,, todate, aetivate much more easily than TiFe as shorlrn in TABLE III.

The percent reaction is a measure ót hydrogen content actuallyattained versus the maximum possible under the conditions speci-fied. It is likely that the same properties which result in anincreased rate of particle attrition, as discussed abover êrêalso responsible for the ease with which such alIoys can beactivated.

Effects of Contaminants on Activitú

The question of the effect of contaminants in hydrogen gasstreams on the absorption-desorption properties of metal-hydro-gen systems is of practical importance for all potential appli-cations of these materials. A portion of the hydride materialsdevelopment program at BNL has been concerned with this problem.The contaminants of interest are these commonly occurring gases:A2, CAZ, N2,

"Z?tg), etc. fn addition chlorine was of interest

because an ongoinf'program at BNL is involved with the develop-ment of a regenerative HCI system for energy storag:e which

-9- ;TRJ,Æ.IR

utilizes a metal hydride as the hydrogen storage medium [¡O-g¿].Results of the poisoning experiments are given in TABLES IV andv. TiFe is seen to be very sensitive to oxygen and even as rit-t1e as 1o ppm o, in hydrogen has a detectable effect on the al-loy activity. Also, higher concentration levels result in fasterpoisoning. the TiFe alIoy, however, could be reactivated bystandard techniques even at the highest 02 levels investigated.

The chlorine contamination experiments show that at 10 ppm thereis no detectable effect over the short term with the manganesealloy but at higher concentrations there is an appreciable lossof activity. Also of importance is the observation that in thelatter case the Famples could not be compretely reactivatedwhich is in opposition to the oxygen findings. Both c12 and o,may have transformed to HCl and H2O under the experimenlal

"onlditions, ho\nrever, whether this waã the case is probabry only ofacademic interest. we did not analyze the gaseous phase or thesample surfaces since our primary concern r^las of a more quali-tative nature. Future experiments should be designed to deter-mine the acÈua1 gas phase species, their concentration levers,and to study the effects of contaminants on the surface of thealloys.

Until more quantitative and fundamental studies are undertaken,as mentioned abover v/ê can only speculate as to mechanisms ofpoisoning. All that can be said nor¡.r is that it appears that ac-tive surface sites for hydrogen absorption-desorpt5-on reactionsare being blocked by the contaminants, oy their reaction prod-ucts such as Hro and Hcl. The formulation of hydrog:en storagealloys which aie not easily poisoned is very desirable but willprobably require an understanding of the mechanism which obtainsat the surface-gas interface.

Desicrn o amum for a eneratiEnercrv Storaqe Svstem

The use of metal-hydrides as energD¡ storage media in conjunctionwith an electrolyzer and fuel cel1 for utility peak shaving ap-plicatíons has been an area of actíve interest at BNL for thepast several years. An alloy with appropriate storage proper-Èies for use with a water electrolyzer-fuer ce1I unit has pre-viously been designed [rz]. The composition of the ternary ar-loy for this application corresponded to TiFeg,7gMng.2g. rmple-mentation of a large-scale unit employing the above components,hcrv¡ever, is not being: pursued because of its unfavorable

JRJ/JJR -10-

economics. A potentially more cost effective and efficientener$y storage system for utility load leveling applications iscurrently under development. The proposed system consists of aGeneral Electric solid polymer electrolytíc ce1l, metal hydridehydrogen storage, and líquíd chlorine/hydrochloric acid étoragesubsystem [gs]. rn the storage mode, hydrogen and chlorine areproduced and stored. rn the discharge mode the hydrogen andehlorine are recombined in the same ceI1 producing electric en-ergy. The waste heat from the recombination reaction is usedto supply heat of decomposition for the hydride. Reference de-sign condítions for the deveropment of a hydride storage sub-system for the unit specify a maximum hydrogen charging pres-sure of 500 psia (34 atm) at 40"c, discharge to 45 psia (3 atm)at 60"c and a storage capacity of >1.5 vtt- yo. rt is also re-quired that the hydride be capable of attaining saturation in aperiod of fíve hours. A number of ternary alloys were formu-lated to meet these ctiteria as shown in TABLES Vf and \rJI. Thealloy with the opÈimum properties, to date, is one having thecomposition TíFeg.g5h0.15. Its pressure-composition character-istics are summariãea iñ-Figs. ro and 11. A point of interestis that the hysteresis for this material is considerably reducedcompared to the titanium iron-hydrogen system. This is signifi-cant since a reduction in the reference design charging pressurewourd lead to a yet larger difference in the maximum hydrogenstorage capacity between fife'.gShO.tS and TiFe.

CONCLUSIONS

The ability to "tailor make" a substituted ferrotitanium alloyhydride with properties matched to a particular storage applica-t,ion extends greatly the versatility of such metar-hydrogen sys-tems for future use. Although much work remains to be d,one indefining and understanding the properties of these potentiallyattractive metal-hydrogen systems, vrê have tried to present herea cross-section of investigations in areas of practical impor-tance which should be considered in any present or future use ofsuch materials"

-11- JR/JJR

ACKNOIüLEDGEMENTS

The authors wish to express their gratitude to s. Bookless, A.Holtz, and J. Hughes for their expert technical assistance inthe laboratory. A1so, we would like to express our thanks toG. Sandrock of the International Nickel compãny for his adviceand assistance in matters of a metallurgical nature.

REFERENCES

1- Reilly, J. 'J. and l{iswall, R. H., Jr. r Formation and prop-erties of Iron Titanium Hydrid.e, Inorg. Chem. 13, zLg (Lg74r.

2. Pebler, A. and Gulbransen, E. A., Equilibrium studies^onthe.rsystems zrcr2-El2t zlv2-H2t and ZrMoz-Hz Between oo and900-C, Trans. TIVIS-AIME, 239, 1593 (1967).

3. Yamashita, T., Gamo, T., Moriwaki, y., and Fukuda, M.,Hydride Formation of Ti-Mn Binary A1loys, J. Jap. Inst. ofMetals, 4L, No. 2, ]4B (1977't.

4. Ishido, Y., Ni-shimiya, N., and Suzuki, 4., preparation andEquilibrium Study on Zirconium Mang:anese Hydride (zrl,mrH*) ,Denki Kagiaku, 45, No. l, 52 (]-g77l .

5. Machida, Y., yamadalza, T., and Asanuma, M., Hydride Forma-tion of cl4-Type Ti Alloys, pres. in part at the rnterna-tional Symposium on Hydrides for Energry Storage, Geilo, Nor-wâ}¡r Augn:st 1977.

6. Sandrock, G. D., A New Family of Hydrogen Storage AlloysBased on the System Nickel-Mischmetal-Ca1cium. Proceedings12th IECEC, 1, 95I, VÍashington, D.C. , L97'7.

7 . Buschow, K.H.J. and van Mal, H. H., phase Relations and Hy-drogen Absorption in the Lanthanum-Nickel system, J. Less-Comrìon Meta1s , 29, 203 (L972, .

8. van Mal, H. H., Buscholrr, K.H.J., and. Ir{iedema, A. R., Hydro-gen Absorption in LaNiU and Related Compoundsi ExperimentalObservations and Their-Exp1anation, J. Less-Conìmon Metals,35, 65 (r9741.

.TRIJ.]R -12-

9. van Vucht, J.H.N., Kuijpers, F. A.r and Bruning' H.C.A.M.,Reversible Room Temperature Absorption of Large Quantitiesof Hydrogen by Intermetallic Compounds, Philips Res. Re-ports 25, 133 (1970)

10. Reilly, J. J., Hvdroqen: Its Technoloqy and Implications,Vol" II, 1st Ed., K. E. Cox and K. D. I^Iilliamson, Eds.,CRC Press, C1eveland, Ohio, L977, Chapter 2.

11. Rei1ly, J. J. r Applications of Metal Hydrides, pres. inpart at the International Symposium on Hydrides for EnergyStorage, Gei-lo, Nonr,ray, August L977.

12. Reilly, .I. J., Hoffman, K. C., Strickland, G., and Wiswall,R. H,, Iron Titanium Hydride as a Source of Hydrogen Fuelfor Stationary and Automotive Applications, 26t-}:, AnnualProceedings of Power Sources Conference, Atlantic City,Ne\^¡ Jersey, ltay !974.

13" Strickland, G.,and Rei11y, J. J., Operating l!Íanual for thePSE&G Hydrogen Reservoir Containing Iron Titanium Hydride,BNL 50421, February L974.

14. VÍol1ey, R. L. and Simons, IIo M., Ilydrogen Storage in Ve-hicles--An Operational Comparison of Alternative Prototypês,Proceedíngs of Fuels and Lr¡bricants Meeting, St. Louis,Missouri, June l-976

1q- Rrrer.hner- TT- andl Sar-rfferer- H- - Resrrlts of Hvdride Research¿¿a t LL. ¡-v-vÈ, ^¡. t

and the Consequences for the Ðeveloprnent of Hydride Ve-hicles, 4th International Slzmposium on Automotive Propul-síon Systems, Arlington, Virginia, April L977.

16. Dangerfield, J., Mass Transit in a Hyd,rogen Economy, HY-d.rogen Progiress, 2nd Quarter 1977, Billings Energ:y Corpora-tion, Provo, Utah.

17 . Rei11y, J. J. and Johnson, J. R., ritanium A11oy Hydrides,Their Properties and Applications, Proceedings of the lstVlorld Hydrogen Energry Conference, Miami Beach, Florida,l,larch L97 6.

-13- JRJ/JJR

18. Reil1y, J. J. andWiswall, R. H-, The Effect of Minor Con-stituents on the Properties of Vanadium and Niobium Hy-drides, Proceedings of the International Meeting on Hydro-gen in Metals, Julich, Germany, Ivlareh L972.

19. Rei11y, J. J. and Wiswa1l, R. H., The Reaction of Hydrogenwith Alloys of Magmesium and Copper, Inorg. Chem. !, 22ZO(1967).

20. Sandrock, G. D., The Interrelations Among composition,Microstructure and Hydridinq Behavior for alloys Based onthe Intermetallic Compound FeTi, Final Report, ContractBNL-35241OS, June 1976.

2L. llurakami, Y. and Enjyo, T., Investigation of Ti RichTi-Fe-¡{n Alloy System on the Solid Equilibrium Relation inthe Ti-Fe-Ivü: System, Nippon Kinzoku Gakkaishí 22-, 329 (1958) .

22. Lundin, C.8., Lynch, F.8., and Magee, C.8., Some UsefulRelationships Between the Physical and Thermodlmamic Prop-erties of Metal Hydrides, Proceedings Ilth IECEC, StateLine, Nevada, September !976.

23" Lund.in, C. 8., Lynch, F. 8., and Magee, C. 8., to bepublished in J. Less-Common Metals-

24. Rosso, M. J., Jr., Strickland, G., and Milau, J. S., pres-sure-Drop-F1uid-F1ow Correlãtion for Fixed Beds of Small,Trrcrrrr I ar'l rz (h:rrar:l Þ¡r{-i n'l ac TltrlT. qfìEeq I\Tnrran}rar 1 o-'7

^:,-__--J¿v¿v¿,

25. Strickland, G., Some Observations on the Effects of theVolumetric Expansion of Iron Titanium on Vessels Built atBù¡L, BNL 23130, August L977.

26. Strickland, G., State-of-th'e-Art Sununary of the TechnicalProblems Involved in the Storage of Hydrogen Vía Metal Hy-dridesr prês. in part at the International Conference onAlternative Eneúg¡¡ Sources, Miami Beach, Florida, Decembert977 -

27. Reilly, J. J., Johnson, J. R., and Reidinger, F., The Reac-tions of Hydrogen with Intermetallíc Compounds, pres. inpart at the 2nd Joint CIC-ACS Conferenee, Montreal, Canada,It{ay-June L977.

,IR,J,/JJR -L4-

28. Reilly, J. J-, Synthesis and Properties of Useful MetalHydrides: A Review of Recent Work at Brookhaven NationalLaborator!t pres. in part at the International Symposiumon Hydrides for Energry Storage, Geilo, No:lray, Augrust L977 "

29. Johnson, J. R. and Reilly, J. J., The Metal Hydride De-velo¡rment Program at Brookhaven National Laboratory, pres.in part at the DOE Chemical Energy Storage and HydrogenEnergy Systems Contracts Review lvfeeting, Hunt ValIey, Mary-land, November 1977.

30. Gileadi, 8., Srinivasan, S., Salzano, F. J. Braun, C.,Beaufrere, 4., Gottesfeld, S., Nuttall, L. J., and LaConti,A. 8., J. Power Sources, in press.

31. Srinivasan, S., Yeo, R. S., and Beaufrere, 4., The Hyd,rogen-Halogen Storage System: Preliminary Feasibility and Eco-nomic Assessment, BNL 22L64, January L977.

32. Beaufrere, 4., Yeo, R. S., Srinivasan, S., McEIroy, J.,and Hart, G., A Hydrogen-Halogen Energv Storage System forElectric Utility Applications, Proceedings 12th IECEC,I,rlashington, Ð. C., 1977.

33. McBreen, J., Yeo, R. S., Beaufrere, A'., Chin, Ð. T., andSrinivasan, S., The Ïlydrogen-Chlorine Energy Storage Sys-tem, pres. in part at the DOE Chemical Energry Storage and.Hydrogen Energi¡¡ Systems Contracts Review lvleeting, HuntValley, It{aryland, November L977 .

34. Reilly, J. J. and Johnson, J. R., Metal Hydride Program atBNl--Current Status and Future Plans, Proceedings of theERDA Contractors'. Review Meeting on Chemical Energ¡¡ Stor-age and Hydrogen Energr¡¡ Systems, Airlie, Virginia, Novem-ber L976.

-15- JRJ/,JJR

T¡IBLE I

PARTTCLE SÏZE ÐTSTRIBIIIION FOR TiFeo.7ho.18 (INCO #1)AITD TiFe* eå A 5{'NqrIoN oF TIIE NU¡{BSR oF

HÐRIÐTNG-DSITYDRIDING CYCLES

W'eiqht Percent

Screen ClassificationMesh

FeTi**5478 Cvcles

11.4

2.92.L4.3

14.7

16.4

11.0

t;:;4.22.37.7

+16

+20

+30

+4A

+60

+80

+100

+140

+200

+275

+325

-325

5.32.A

1.510.1

32.2

19.3

17"1

l_4.1

6.40.6

2.Lt2.425.5

2L.5

11.8'to .7

*Commercially obtained as a 500 lb. ingot.**Initial particle size range -LA/+16 (U.S. Standard Sieve).

Ti'eo. zho . rB**5 Cvcles 1115 Cvcles

,JR.I/JJR -16-

Ip{I

ACTIVITY OF TireAS A TTTNCTTON OF

A1loyCompo$ítion

TiFe*

TiFeo. zgho . rs(rNco #2)

TíFeo.76ho.14

CJfrC{

qt-iN

A¡ilD TiFe TvürxyTTÍE NU¡{BER OF

Number ofC.ycl-es

19600

28400

2 6300

TABLE TT

*Commercíally obtaíned as a 500 lb. ingot.

TemperaturechgnSe,/Cycle

C,-Avq.

-1 to +100

-1 to f100

-1 to +100

ALLOYS IN A ÏI2 EI{VTRON¡4Ehn

TTYDRTD TNG-DEHYDRTDING CYCLES

Pressurechange/CycLePSIA, Avg.

428-459

T79-2LI

87-ILA

CompositionChange,/Cycle_ a (Hlri) -

I.1Il.0g

0. 96

t-tñCt

qTJH

TABLE TTT

LOVI TEMPERATURE ACTTVATTON EXPERTMENTS FoR TiFe I{n ALLoYsxy

AlloyComposition

TíFE*

ríreo.9oho. ro

IHæI

rireo.gsho.15

TíFeo. goho.2o

TiFeo.7oho.lB(rNco #1)

ríFeo " 7oho.2o

TiFeo.932ho .agr(rNco #5)

TíFeo.goho.096(rNco #4)

TiFeo " 77h0. 09

TiFeo.73ho.13

Outgass ingTemperature

io"'t100

i,s

25

25

25

50

25

25

EJ-apsed TímeAt 500 PSIA (hr)

72

27

6

45

25

3.5

68

19

f ina 1 e/o

*Commercially obtaíned as

ri/r,4

0

4.24

0.97

0.58

0.91

0.94

0.81

0.86

0.98

0. gg

Reactíon0

25

100

60

91

94

100

96

100

100

25

25

27 .5

19

a 500 Ib. ingot.

TABLE IV

EFFECT OF 02 IN H2 ON THE ACTTVITY OF

TiFe-H *x

rnitial o, Percent Reduction rnContent (ppm) Composition Chanqe À (¡tlfiì

5 Cycles 40 Cycles

2910

100

1000

11000

6

4A

20

BO

90 90

:tCommercially obtained as a 500 lb. ingot.

TABLE V

EFFECT OF C12 IN H2 GAS ON THE ACTTVITY

oF TiFeo.9ho.l-Hx

rnitial cI, Activity Reactivatio4 ActivityContent (ppm) foss (d Temperature (oC) Recovered (%)

10

100

1000

1000

50

75

75

110

1I0

300

400

88

63*

75*

75*

*These are maxjmum values. After exposure to high C12concentrations there was a slow deterioration inperformance after reactivation.

-19- JRJ,/JJR

TABLE VT

ITYDROGEN STORAGE CAPACITIES FOR TiFe JvfTT ALLOYSxyI'¡IDER PROiTECTED H2-C12 CELL OPERÃ,TII{IG COIqDITIONS

AlloY lvlaximum H,

Composition Storage Capacitv Tft %

TiFe* 1.55rireo.goho.096 1'50

(rNco #4)

tif e'.90h0.l' l-.65

TiFeo.B5ho. 15 1'80

TiFeo.gOho.Zo 1.60

TiFeO.7Oh'. 30 O .90

TiFeo.932ho.og4 1'35(rNco #5)

*Commercially obtained as a 500 lb. ingot.

TABLE VTT

CHARGTNG TESTS OF 5 HOÜRS DÎ'RATTON FORCOMPLETELY ACTI\ZATEÐ TiFex¡4nv ALLOYS AT

500 PSIA emo 250c

A1loy Final %Composition Éî/NI Reaction

Tireo,93ho .ag4 o '81 1oo

(rNco #5)

TiFeO.BshO. 15 0 .98 100

iTRJ/J.]R -20-

4 - NEGATIVE NUMBERS: NEGATIVE NUMBÊRS FOR FIGURES AR6I

,FIG. NO. NEG. NO. FIG. NO. NEG. NO.

t 1:6-,14:73 ..6 : ..1?:.F.ø-.7F r,r? rl:914:77 ,7 t2-g7-7s l

9 . 3--.?67:7..6... B ., +t;,8+6 _j,lz4 . -12:5-5175-.. ....9-. ,,

' r-324-76ã 12-:45.:7tr. Lo r-r_o;157r _77

.

NEG. NO. FIG

to-1572-77

(\'TEË lolrjæ.:)(nCNlrjE,fL

z.It--

(Jocn¡õ

too

o.4 0.6 0.9AîOM RATIO H/{Fe

F'ig. 1. Pressure-CoÍposition Isotherms forRatios at 40oc: {e} 60.5 wt % Fe,wt % Fe, 49.2 wt % Ti; (c) 36.? wt

l.o+T i)

Alloys of39.5 wt %/" 7e, 63.2

1.2

Various fe,/fiTi ; tB) 50 .5ï"t % Ti.

.F^\

L l- ll\ E-u )RI

.to

.t5

.20

.30

¡il

ffi

,/''

'-':::.r/ I

IE

fìeflsnsno

(\¡

-EËtotrjæ.:)aØl!E.o_

z.IF()o8toõ

o.r o.3 0.5 0.7 0.9H /(Ti+ Fe + Mn )

Pressure-CCImposition Desorption Isothermsof Various l"ln Contents at ãOoc.

for FeTi A11oysFig.2.

VACUUMH2

PRESSURE

CALIBRATED VOLUME

JACKETEDREACTOR

Fíg. 3. Schematic of a Station on the }Iigh Cycle Test Apparatus.

PRESSURETRANSDUCER

HEAT TRANSFER FLU¡D

GlT to=t-

l¡¡fr:)ûcDl¡JË,o-

zoÞsC)oa,v,ol

A Comparison ofand TiFeo.gcoCI.l

Ti FeTi Fe., Co.,

40 0c

.5 .6 .7 .8 .gHlM

40oc Ðesorption Isotherms-H Systemg.

Fig. 4. for the TiFe-H

roo

t',

?

/

I.--/ /

I

C\l

!to

=þ

lr,fts:f,Ø(t,lr¡É,fL

zotrfoooaôl

I

Ti Fe

Ti Fe.e Gr.,

- T¡ F"."u Cr.os

40 "c

.5 .6H/M

Fig. 5. A Comparison of 4OoC Ðesorption ïsotherms for the TiFe-H,Tife'.9CrO.l-H and tiF"O.gSCt..OS-II Systems

.3.2

q¡:E

=t-lrjË,:fÎnU'l¡JEo-

zot-s()oU'C"

ô

Ti Fe

Ti Fe Gu.,

40 0c

--o. .l .2 .3 .4 .5 .6 .7 .g .g l.O l.lH/M

rig. 6. A Comparison of 40oc Ðesorption Isotherms for theTiFe-H and TiFecuO.l-H Systems.

o¡T

=l-

l¡Jfr:)aCf'l¡JÉ.fL

z.9t-ÍooU)(t)

ô

o.

Fig. 7.

_Ti Fe

'Ti Fe., Mo.,

40 "c

.t .2 .3 .4 ,5 .6 .7 .8 .9 l.o l.lH/M

A comparison of 4ooc Ðesorption ïsotherms for the TiFe-Hand TiF"'.9Mo0.1-H Systems.

GI

TE r/^

t\.,olijE.

=U)alrJE.o_

z.al-õ8 ro(nõ

roo

o.r o.3 0.5 0.7 0.9 l.lH/(Ti +Fe+Mn)

Fig' 8. A Comparison of Pressure-Composition Desorption^Isothermsfor .Annealed and "As-Cast', TiFeO.Z6hO_13 ãt 40oc-

o.tt.3

Jåd

ËFaå{ïoö

IIo

I/!

,ij40"

.-o¿t/ -"r/

t/tto's-'o-o'o'

,/" /io/

/ ^.o/Lo-o'o-e

O ..AS CAST..

O ANNEALED

/"-"-7

l/'I

III

Alr ro

=F

¡¡,æ,fv,v,lrlorfL

=:)ËfD

J

=o¡rJ I

t,

- --t /

.5 .6H/M

-T¡

Fe.7Mn2

30 0c

+ SORPTION

+- DESORPTION

Fig. g. Hysteresis in FeTi and, TiFeO.7h*.Z at 30oC.

(\¡

-EElolJ.iÉ.=(ncnl¡JÉ.o-z.o[-

õoØar-¡ I

o.? oA 0.6 o.8 t.o t.2

ATOM RATIO H/(Ti +Fe+Mn)Fig. 10, Pressure-composition Ðesorptj-on rsotherms for the

fiFe'. g'h'.l5-II System at Several Temperatures.

60"

40"

toô¡

-

€oIllE.:)Ø(nLJE.o_

=)É.m !^-i l.u

=øl¡J

roo

a

a

40"

-<

o.l I | ¡ I I I I I I I I I I I I Io.2 0.4 0.6 0.8 t.o t.2 t.4

ATOM RATIO H/ff. + Fe + Mn)

Fig. 11. Hysteresis for the TiFeo.gsho.l5-Ir system at 40oc.