Flux growth utilizing the reaction between flux and crucible

3.6 MB·English

by  J. -Q. Yan

Flux growth utilizing the reaction between flux and crucible
J.-Q. Yan1,2
1Materials Science and Technology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA
2Department of Materials Science and Engineering,
University of Tennessee, Knoxville, Tennessee 37996, USA
(Dated: January 16, 2015)
Fluxgrowthinvolvesdissolvingthecomponentsofthetargetcompoundinanappropriatefluxat
hightemperaturesandthencrystallizingundersupersaturationcontrolledbycoolingorevaporating
theflux. Arefractorycrucibleisgenerallyusedtocontainthehightemperaturemelt. Thereaction
between the melt and crucible materials can modify the composition of the melt, which typically
5 resultsingrowthfailure,orcontaminatesthecrystals. Thusoneprincipleindesigningafluxgrowth
1
is to select suitable flux and crucible materials thus to avoid any reaction between them. In this
0
paper, we review two cases of flux growth in which the reaction between flux and Al O crucible
2 3
2
tunes the oxygen content in the melt and helps the crystallization of desired compositions. For

n the case of La5Pb3O, Al2O3 crucible oxidizes La to form a passivating La2O3 layer which not only
a prevents further oxidization of La in the melt but also provides [O] to the melt. For the case
J of La Na Fe As , it is believed that the Al O crucible reacts with NaAsO and the reaction
0.4 0.6 2 2 2 3 2
5 consumes oxygen in the melt thus maintaining an oxygen-free environment.
1
Keywords: A2. Growth from high temperature solutions, B2. Superconducting materials, B1. Arsenates,

A1. Solubility,A1. Diffusion
]
i
c
s
Introduction In many cases, more than one kind of flux can medi-
-
l ate the crystal growth of one desired compound. Crys-
r
t tals of the same nominal composition but grown out of
m Materialssynthesisandcrystalgrowthinmoltenfluxes
different fluxes might exhibit different physical proper-
. are compelling routes to novel materials with intrigu-
t ties,whichcouldleadtoconfusionsinunderstandingthe
a ing physical properties important for both scientific re-
intrinsic properties. One good example is the physical
m search and technological applications. Flux growth is
propertiesofK-dopedBaFe As ,oneimportantmember
- alsoknownashightemperaturesolutiongrowth;theflux 2 2
d ofFeAs-basedsuperconductors. Kimetal[13]studiedthe
melts at high temperatures and acts as the solvent in a
n low temperature specific heat of Ba K Fe As single
solution growth. In a typical flux growth, first the com- 0.6 0.4 2 2
o crystals grown out of Sn, In, and FeAs fluxes, and found
c ponents of the target compound are dissolved in a flux,
that the physical properties depend on the preparation
[ which is usually molten salt, oxide, or metal inside of a
method. Thisillustratestheimportanceofgrowingcrys-
refractorycrucible. Then, crystalgrowthtakesplaceun-
1 talsoutofanappropriateflux. Therearesomeimportant
v der supersaturation controlled by cooling or evaporating
factors that have to be considered in selecting the right
0 theflux. Bydissolvingthecomponentsofadesiredcom-
flux.[1, 6, 14] One of them is that the reaction between
7 pound in a flux with a low melting temperature, some-
7 flux and crucible should be negligible.
times in controlled atmosphere or sealed ampoule, flux
3
growth is good for the growth of a large variety of ma-
0 In a flux growth, a refractory crucible is needed to
. terials, such as compounds that would sublimate, com- hold the melt at high temperatures. Depending on the
1
poundswithahighmeltingtemperature,volatilecompo-
0 composition of the desired compound, the properties of
nents,poisonouscomponents,oraphasetransitionbelow
5 flux, and the temperature profile, crucible materials can
1 the melting temperature. be metals (Fe, Ni, Ta, W, Mo, Pt), quartz, graphite,
:
v Flux growth of oxides has been well developed in the nitrides (BN), or oxides (Al2O3, MgO, ZrO2). At high
i lastcentury. Agoodtextbookaboutthistopicis”Crys- temperatures, a serious reaction between flux and cru-
X
tal Growth from High-Temperature Solutions” by D. El- cible can destroy the crucible and ruin the growth. A
r
well and H. J. Scheel.[1] There are also some good re- minor reaction can modify the composition of the melt,
a
view articles. [2–4] Crystal growth and materials explo- whichleadstothefailureofthegrowth,or,inmostcases,
ration from metallic flux were mainly developed in re- contaminates the crystals as a substitution or an inclu-
centyearswithamajorcontributionfromP.C.Canfield sion. Crystal contamination from crucible materials can
and Z. Fisk.[4–12] Compared to the flux growth of ox- significantly affect the physical properties. This is well
ides which is normally performed in air or flowing gas illustrated by the growth of YBCO superconductors in
(O ,N ,orAr),growthofintermetalliccompoundsfrom an Al O crucible. The BaO-CuO flux attacks Al O
2 2 2 3 2 3
metallicfluxor saltismorechallengingbecausemostin- crucibles and Al O dissolved into the flux contaminates
2 3
termetallics tend to oxidize in air and some are moisture the crystals as substituent that suppresses the supercon-
sensitive, volatile, and/or poisonous. ducting transition temperature by over 20K.[15] Thus,
2
LaO
2 3
LaPbO
5 3
FIG.2: (coloronline)SEMpictureofLa O layer. La Pb O
2 3 5 3
FIG.1: (coloronline)La-CophasediagramfromASMAlloy nucleates on La O layer.
2 3
Phase Diagram Database.[17]
Ce Pb O single crystals.[16] Once Co is absent in the
5 3
whendesigningafluxgrowth,agoodcombinationofflux starting materials, no Ce Pb O crystals could be ob-
5 3
and crucible is normally selected so that no or negligible tained. This, together with the fact that no Co contam-
reaction between flux and crucible is expected at high ination was found in the as-grown crystals, suggest that
temperatures. Co is an important component in forming a proper flux
In this paper, we provide a different view of the reac- mediatingthegrowthofCe Pb O.Thesuccessfulgrowth
5 3
tion between flux and refractory crucible materials: in of Co-free La Pb O in our study suggests a similar role
5 3
somespecialcases,thereactioncanbeemployedtofacil- of Co in the growth. Investigation of the La-Co and La-
itatethegrowthofdesiredcompositions. Specifically,we Pb binary phase diagrams suggests that Co is important
reviewthefluxgrowthofLa5Pb3OandLa0.4Na0.6Fe2As2 inloweringthemeltingtemperatureofthemixture. Fig-
in which the reaction between flux and Al2O3 crucible ure1 shows the La-Co binary phase diagram.[17] In a
helps control the oxygen content in the melt for the wide composition range 10<x <55%at, La Co al-
Co 1−x x
growth of desired crystals. loy melts below 850oC. Even though Pb melts at a low
temperature of 320◦C, the melting temperature of La-
PballoysincreasesrapidlywhenintroducingLaintoPb
Crystal Growth of La5Pb3O melt; the line compound La5Pb3 melts congruently at
1450◦C.
La5Pb3O single crystals were grown out of a mixture Itisinterestingtonotethatnooxygenwasintroduced
of La, Co, and Pb. Starting materials La (99.999% inthestartingmaterialsbuttheresultingcrystalscontain
Ames Laboratory), Co (99.999%, Alfa Aesar), and Pb oxygen. After decanting, a thin layer of La O was al-
2 3
(99.9999%, AlfaAesar)wereplacedina2mlAl2O3 cru- ways observed inside of the alumina growth crucible and
ciblewithamolarratioof7:2:1. Acatchcruciblecontain- it takes the shape of the crucible. Figure2 shows a SEM
ing quartz wool was mounted on top of growth crucible pictureofpartoftheLa O layer. Thehighpurestarting
2 3
and both were sealed in a silica ampoule under approx- materials are not expected to introduce that much oxy-
imately 1/3 atmosphere of high pure argon gas. The gen. The growth was performed in a sealed quartz tube
sealed ampoule was heated to 1150◦C, stayed at 1150◦C with the highest homogenizing temperature of 1150oC
for 4 hours, and then cooled to 850◦C over 60 hours. which is below the softening temperature of quartz. So
Once the furnace reached 850◦C, the excess flux was de- it is unlikely that significant amount of oxygen diffuses
canted from the rectangular crystals with the typical di- into the quartz tube during crystal growth. As shown in
mension of 0.5×0.5×2-4mm3. Room temperature x-ray Figure2,somequartzwoolinthecatchcruciblefalldown
powderdiffractionofpulverizedsinglecrystalsconfirmed intothemelt, whichmightprovideoxygentooxidizeLa.
single phase of as-grown crystals. The stoichiometry of However, in a growth performed without using quartz
as-grown crystals was confirmed with elementary anal- wool inside of the catch crucible, both the La O layer
2 3
ysis as well as single crystal x-ray and neutron diffrac- and La Pb O single crystals were obtained. Therefore,
5 3
tion performed at room temperature. Detailed crystal the alumina growth crucible, which has direct contact
characterization and physical properties will be reported with the melt, is expected to be the only source for oxy-
elsewhere. gen in La Pb O crystals and the La O layer. As a rule
5 3 2 3
A similar procedure was previously employed to grow of thumb, rare earth concentrations over 12% in a wide
3
Quartz tube
quartz wool
Ta tube
(b)
(a)
LaFeAsO
2[La] + 3[O] = LaO
2 3
FIG. 3: (color online) Schematic picture illustrating the
La O layer formed at high temperatures.
2 3
variety of intermetallic fluxes react with Al O .[18] In
2 3
the growth of La Pb O, there are 70%at La in the melt
5 3
and oxidization of La metal in the melt by the growth
crucible is expected. This oxidization leads to the for-
mation of La2O3 layer and also explains why the La2O3 (c)
layer takes the shape of the growth crucible.
As shown in Figure2, the rectangular La5Pb3O grains FIG. 4: (color online) (a) Growth ampoule. For the growth
nucleate on the surface of La O layer. We believe that of LaFeAsO, all starting materials were loaded in the Ta
2 3
theoxidizationlayeriscriticalforthegrowthofLa5Pb3O tube. For the growth of La0.4Na0.6Fe2As2, the starting ma-
crystals: (1)itservesasthesourceofoxygentothemelt terials were held in an Al2O3 crucible inside of Ta tube.
(b) Temperature profile for both growths, (c) Schematic
and controls the oxygen content in the melt by the re-
pictures illustrating the growth of LaFeAsO in Ta crucible
action 2[La]+3[O]=La O (s); (2) it also serves as the
2 3 and La Na Fe As in Al O crucible. Crystal picture of
0.4 0.6 2 2 2 3
passivating layer to prevent further oxidization of rare
LaFeAsO is from ref[20] and that of La Na Fe As from
0.4 0.6 2 2
earth metal. The La/Co ratio in the starting materials ref[19].
stays close to the eutectic with the melting temperature
about640oC(seeFigure1). OxidizationofLametalwill
modifythecompositionofthemelt. Themeltingtemper- La O passivating layer. With a similar procedure, we
2 3
ature of the melt increases when the flux becomes more have been able to grow R Pb O (R=La, Ce, Pr, Nd,
5 3
Co-rich, and in the worst case La-Co binary compounds and Sm) crystals. However, we failed to grow R Pb O
5 3
mightprecipitate. Adecantingtemperatureof850oCwas crystals with heavier rare earth ions even though a rare
selected considering the composition change of the melt earth oxide layer also forms. This might signal limited
duetotheoxidizationofLaduringthegrowth. Thelow- oxygensolubilityand/ordiffusivityinR-Co-Pbmeltwith
est decanting temperature used in our growths is 825oC. heavy rare earth ions.
The successful decanting at 825oC suggests that La:Co
ratio should be less than 55%at without considering the
effect of alloying with Pb on the melting temperature of Crystal Growth of La0.4Na0.6Fe2As2
La-Co binary alloy.
Figure3 illustrates how the reaction between growth The growth and physical properties of
crucible and flux helps the growth of La Pb O: at high La Na Fe As single crystals were reported
5 3 0.4 0.6 2 2
temperatures, Al O oxidizes La metal forming La O elsewhere.[19] When looking for proper crucible materi-
2 3 2 3
passivating layer which controls the oxygen content in als for the growth of LaFeAsO superconductors out of
the melt; La Pb O crystals nucleate on the surface of NaAsflux, it wasnoticedthatLa Na Fe As crystals
5 3 0.4 0.6 2 2
4
werealwaysobtainedonceAl O crucibleisusedtohold hole asymmetry without disturbing the FeAs conduct-
2 3
the melt but LaFeAsO crystals result when Ta crucible ing layer. There are other 122 compounds with al-
is used, even though we started with the same starting kali metal, such as AFe As (A=Alkali metal),[22–24]
2 2
materials and used identical heat treatment process. AE Na Fe As (AE=alkaline earth ion).[25–27] Crys-
1−x x 2 2
The difference is illustrated in Figure4. For the growth talgrowthof122compoundswithalkaliionsisnormally
of LaFeAsO single crystals, the starting materials of a performed using a similar growth setup with an Al O
2 3
homogeneous mixture of LaAs, (1/3Fe+1/3Fe O ) or crucible, a Ta tube, and a quartz tube (see Fig 4(a)).
2 3
FeO,andNaAswiththemolarratioof1:1:15wasloaded The quartz tube protects the Ta tube from oxidization,
into a Ta tube.[20] For the growth of La Na Fe As andTatubepreventstheevaporationlossofactivealkali
0.4 0.6 2 2
single crystals, the same homogeneous mixture was metals and reaction with quartz tube. At high temper-
loaded into an Al O crucible, which was further sealed atures, attack of Al O crucible by alkali metals takes
2 3 2 3
inside of a Ta tube under 1/3 atmosphere of argon gas. place leading to a variation of the crystal composition
The use of Ta tube in both cases is necessary to protect fromthenominalone. Buttothebestofourknowledge,
quartz tube from NaAs at high temperatures. thereisnoreportofoxygencontaminationresultingfrom
As shown in Figure4, the only difference between the the reaction. For the growth of other 122 compounds
above two growths is the use of an Al O crucible which without volatile alkali ions, little reaction was observed
2 3
has a direct contact with the flux. The dramatic dif- between flux and Al2O3 growth crucibles.[28] The reac-
ference in the crystal composition forces us to consider tion observed when growing La0.4Na0.6Fe2As2 suggests
the possibility that the Al O crucible absorbs oxygen if a low-melting compound containing oxygen forms, like
2 3
introduced by iron oxide into the melt. The formation NaAsO2 in NaAs flux, oxide crucible might not be the
of NaAsO has been proposed to account for the solu- suitablecontainerforthegrowthofoxygenbearingcom-
2
bility of oxygen in an NaAs flux. NaAsO has a low pounds. However, this mechanism could be employed
2
melting temperature of about 600oC, which is similar to to eliminate oxygen introduced by starting materials to
that of NaAs.[21] After crystal growth, the Al O cru- improve crystal quality.
2 3
cible turns to be grey and reaction between the crucible
and flux was always observed. Sometimes, the Al O
2 3
growth crucible cracks after growth. Residual flux was Summary
foundbothinsideandoutsideofthegrowthcrucible. Un-
fortunately, we did not further investigate the reaction In summary, we reviewed the flux growth of La Pb O
5 3
mechanismforthefollowingtworeasons: (1)inatypical andLa Na Fe As singlecrystalsandillustratedhow
0.4 0.6 2 2
growth, about 0.10 gram Fe O was used in a 3 gram the reaction between flux and Al O crucible controls
2 3 2 3
mixture of charge and flux. This can lead to the forma- oxygen content in the melt and the resulting crystals.
tion of 0.12 gram NaAsO . With that little amount of For the case of La Pb O, Al O crucible oxidizes La
2 5 3 2 3
NaAsO , it’s difficult to find the reacting product in the metalinthemelt;thisreactionformsaLa O shellwhich
2 2 3
ampoule; and (2) sodium arsenate hydrate, which is poi- servesastheoxygensourcetothemeltandapassivating
sonous and can lead to stomach pain, nausea, decreased layer which protects the remaining La metal in the melt.
blood pressure, headache, and possibly carcinogenic and For the case of La Na Fe As , the reaction between
0.4 0.6 2 2
teratogenicproblems,mightformandstayonthesurface NaAsO in the flux and Al O crucible helps maintain
2 2 3
of growth crucible when washing the crystals out of flux. an oxygen free environment to facilitate the growth of
So we tried to reduce any contact with the crucible and La Na Fe As . These two growth examples demon-
0.4 0.6 2 2
residual chemicals. Even without any direct observation strate that the reaction between flux and Al O crucible
2 3
of the reacting product, it is reasonable to believe that can be employed to assist crystal growth, especially for
NaAsO reacts with alumina crucible at high tempera- growthsduringwhichtheoxygencontentinthehightem-
2
tures, considering the color change and cracking of the perature melt should be controlled. This provides a new
Al O growthcrucible,andvariousquaternaryphasesin way to look for a proper flux for the growth of some
2 3
Na-Al-As-O system. This reaction consumes oxygen in special intermetallic compounds bearing oxygen, such as
the melt and maintains an oxygen free environment. In RFeAsO superconductors.
contrast, the successful growth of LaFeAsO single crys-
tals from NaAs flux in a Ta tube suggests that the melt
cannot oxidize Ta tube at the growth conditions. Acknowledgment
La Na Fe As isanewThCr Si -type(’122’)com-
0.4 0.6 2 2 2 2
pound with La3+ and Na+ ions occupying the same JQY thanks P. C. Canfield, A. F. May, R. W. McCal-
crystallographic site. From a simple electron counting, lum, and M. A. McGuire for helpful discussions and im-
varying the La/Na ratio in La Na Fe As can proving the manuscript. Work at ORNL was supported
0.5−x 0.5+x 2 2
tune the material to be either electron-doped (x<0) or by the US Department of Energy, Office of Science, Ba-
hole-doped (x>0), which allows the study of electron- sic Energy Science, Materials Sciences and Engineering
5
Division. Part of the growth of La Na Fe As was Crystal Growth, 91, 308 (1988).
0.4 0.6 2 2
performed at Ames Laboratory. [16] R.T.Macaluso,N.O.Moreno,Z.Fisk,J.D.Thompson,
andJ.Y.Chan,ChemistryofMaterials,16,1560(2004).
[17] Okamoto H., Co-La (Cobalt-Lanthanum),
Binary Alloy Phase Diagrams, II Ed.,
Ed. T.B. Massalski, 2, 1201 (1990).
http://www1.asminternational.org/AsmEnterprise/APD,
[1] D. Elwell, and H. J. Scheel, Crystal growth from high-
ASM International, Materials Park, OH, 2006.
temperature solutions, Academic Press INC. (London)
[18] Notes from P. C. Canfield at Ames Laboratory.
LTD.(1975).
[19] J.-Q. Yan, S. Nandi, B. Saparov, P. Cermk, Y. Xiao, Y.
[2] B. M. Wanklyn, Journal of Materials Science, 7, 813
Su, W. T. Jin, A. Schneidewind, Th. Brckel, R. W. Mc-
(1972).
Callum,T.A.Lograsso,B.C.Sales,andD.G.Mandrus,
[3] S. J. Mugavero III, W. R. Gemmill, I. P. Roof, H. C.
Phys. Rev. B 91,024501 (2015).
zur Loye, Journal of Solid State Chemistry, 182, 1950
[20] J.-Q.Yan,S.Nandi,J.L.Zarestky,W.Tian,A.Kreyssig,
(2009).
B.Jensen,A.Kracher,K.W.Dennis,R.J.McQueeney,
[4] I. R. Fisher, M. C. Shapiro, and J. G. Analytis, Philo-
A. I. Goldman, R. W. McCallum, and T. A. Lograsso,
sophical Magazine, 92, 2401 (2012).
Appl. Phys. Lett., 95, 222504 (2009).
[5] P.C.Canfield,andZ.Fisk,PhilosophicalMagazine,65,
[21] J.-Q. Yan, B. Jensen, K. W. Dennis, R. W. McCallum,
1117 (1992).
andT.A.Lograsso,Appl.Phys.Lett.,98,072504(2011).
[6] Z. Fisk, and J. P. Remeika, Handbook on the Physics
[22] M. Rotter, M. Pangerl, M. Tegel, and D. Johrendt,
and Chemistry of Rare Earths, Vol. 12, edited by K. A.
Angew. Chem. Int. Ed. 47, 7949 (2008).
Gschneidner, Jr., and L. Eyring (Amsterdam: Elsevier).
(1989). [23] S. Peschke, T. Strzer, and D. Johrendt, Z. Anorg. Allg.
[7] C.Petrovic,P.C.Canfield,andJ.Y.Mellen,Philosoph- Chem. 640,830 (2014).
ical Magazine, 92, 2448 (2012). [24] A. F. Wang, B. Y. Pan, X. G. Luo, F. Chen, Y. J. Yan,
[8] Y.Janssen,M.Angst,K.W.Dennis,P.C.Canfield,and J.J.Ying,G.J.Ye,P.Cheng,X.C.Hong,S.Y.Li,and
R. W. McCallum, Review of Scientific Instruments 77, X. H. Chen, Phys. Rev. B 87,214509 (2013).
056104 (2006). [25] K.Zhao,Q.Q.Liu,X.C.Wang,Z.Deng,Y.X.Lv,J.L.
[9] X. Lin, S. L. Bud’ko, and P. C. Canfield, Philosophical Zhu, F. Y. Li, C.Q.Jin, Phys. Rev. B 84,184534 (2011).
Magazine, 92, 19 (2012). [26] S. Avci, J. M. Allred, O. Chmaissem, D. Y. Chung,
[10] P. C. Canfield, and I. R. Fisher, Journal of Crystal S. Rosenkranz, J. A. Schlueter, H. Claus, A. Daoud-
Growth, 225, 155 (2001). Aladine, D. D. Khalyavin, P. Manuel, A. Llobet, M. R.
[11] W. A. Phelan, M. C. Menard, M. J. Kangas, G. T. Mc- Suchomel,M.G.Kanatzidis,andR.Osborn,Phys.Rev.
Candless, B. L. Drake, and J. Y. Chan, Chemistry of B 88,094510 (2013).
Materials, 24, 409 (2012). [27] YanpengQi,LeiWang,ZhaoshunGao,XianpingZhang,
[12] M. G. Kanaztzidis, R. Pottgen, and W. Jestschko, Dongliang Wang, Chao Yao, Chunlei Wang, Chengduo
Angew. Chem. Int. Ed., 44, 6996 (2005). WangandYanweiMa,NewJournalofPhysics14,033011
[13] J. S. kim, E. G. Kim, and G. R. Stewart, J. phys.: Con- (2012).
dens. Matter 21, 252201 (2009). [28] S. Aswartham, C. Nacke, G. Friemel, N. Leps, S.
[14] Handbook of Crystal Growth 2, Bulk Crystal Growth, Wurmehl, N. Wizent, C. Hess, R. Klingeler, G. Behr,
Part A: Basic Techniques,Edited by D. T. J. Hurle, S. Singh, B. Buchner, Journal of Crystal Growth, 314,
North-Holland(1994). 341 (2011).
[15] C. N. W. Darlington, and D. A. O’Connor, Journal of