Synthesis of Mg(B$_{1-x}$C$_x$)$_2$ Powders
a r X i v :c o n d -m a t /0507151v 1 [c o n d -m a t .s u p r -c o n ] 6 J u l 2005Synthesis of Mg(B 1?x C x )2Powders
R.H.T.Wilke,S.L.Bud’ko,P.C.Can?eld,and D.K.Finnemore Ames Laboratory US DOE and Department of Physics and Astronomy,Iowa State University,Ames,IA 50011S.T.Hannahs National High Magnetic Field Laboratory,Florida State University,1800E.Paul Dirac Drive,Tallahassee,Florida 323101Introduction Superconductivity in MgB 2near 40K [1]has attracted much interest due to its potential for applications in the 20-30K range.The low upper critical ?eld [2]and high anisotropy ratio,γ=H ab c 2/H ⊥ab c 2[3,4,5],of pure MgB 2,limit
its potential usefulness [6].Carbon doping of MgB 2has been shown to be an e?ective method for enhancing H c 2[7,8,9]while simultaneously decreasing the
anisotropy ratio by increasing H ⊥ab c 2
more rapidly than H ab c 2[10,11].Carbon doped MgB 2?laments require high reaction temperatures which lead to large grain sizes and poor J c values,in spite of the enhancements in H c 2[12].Re-ports of successful fabrication of superconducting wire using powder-in-tube processing [13,14],coupled with the fact that J c values in powder samples can readily be increased by the addition of various particles [15,16,17,18]motivate
a desire for synthesizing single phase carbon doped MgB2powder.Carbon doped bulk polycrystalline MgB2with approximately10%carbon incorpora-tion has previously been synthesized by mixing elemental Mg and the binary B4C[19,20].Systematic carbon doping of Mg(B1?x C x)2with x<0.10has been achieved in single crystals[8,21]and polycrystalline wires fabricated by chem-ical vapor deposition(CVD)[7,12].In this paper we explore the possibility of preparing Mg(B1?x C x)2with x<0.10using B4C as the carbon source.
2Experimental Methods
Powder samples of Mg(B1?x C x)2were prepared in a two step process.First, stoichiometric mixtures of distilled Mg,elemental B,and the binary compound B4C were reacted for48hours at1200o C.The resultant sample was then re-ground in acetone,pressed into a pellet,and re-sintered for an additional48 hours at1200o C.Two di?erent batches were made.The?rst with0.995pu-rity B(metals basis)from Alpha Aesar,and the second with0.9997purity isotopically enriched11B from Eagle Picher.The three main impurities in the 0.995purity B are C,Si,and Fe,which have relative atomic abundances of 0.25%and0.20%,and0.10%respectively.The isotopically enriched11B con-tained0.02%Ta,0.001%Cu,and0.001%Fe.Samples made with isotopically enriched11B were shown to have a higher residual resistivity ratio(RRR)than samples made with nominal0.9999purity B,indicating the isotopic enrich-ment process yields perhaps the purest boron available[22].B4C from Alpha Aesar was used as the carbon source in both runs.The B4C had a nominal purity of0.994metals basis with the two primary impurities being Si and Fe, which occur in relative abundances of0.37%and0.074%respectively,values similar to those in the0.995purity B.Since impurities in the boron have been shown to suppress T c[22],two di?ering boron purities were used to examine the e?ects of the starting boron purity on the T c and H c2(T=0)values in carbon doped MgB2.We found it necessary to use multiple reaction steps to incorporate the carbon as uniformly as possible.To avoid confusion regarding the meaning of x in Mg(B1?x C x)2,we will henceforth refer to the nominal carbon content as x n and the inferred carbon content after the m-th reaction step as x im.
Powder x-ray di?raction(XRD)measurements were made at room temper-ature using CuKαradiation in a Rigaku Mini?ex Di?ractometer.A silicon standard was used to calibrate each https://www.wendangku.net/doc/0c10213347.html,ttice parameters were deter-mined from the position of the MgB2(002)and(110)peaks.DC magnetization measurements were performed in a Quantum Design MPMS-5SQUID mag-netometer.Transport measurements were done using a four probe technique, with platinum wires attached to the samples with Epotek H20E silver epoxy. Resistance versus temperature in applied?elds up to14T were carried out in
a Quantum Design PPMS-14system and resistance versus?eld was measured up to32.5T using a lock-in ampli?er technique in a resistive DC magnet at the National High Magnetic Field Laboratory in Tallahassee,Florida.
3Results and Discussion
Using the lower purity boron,we reacted a sample with nominal carbon con-tent of x n=0.05for48hours at1200o C.The carbon content can be estimated by the shift of the x-ray(110)peak position relative to that of a nominally pure MgB2sample made under the same conditions[7,20].Although the B4C is a di?erent B source than that used for the reference sample,the0.994pu-rity B4C and0.995purity B contain similar concentrations of impurity phases, and we cautiously proceed with estimates of the carbon content ignoring mi-nor di?erences between boron sources.Indexing of the(110)peak for the pure sample and that containing a nominal carbon content of x n=0.05yielded an inferred carbon level after this?rst reaction step of approximately x i1=0.031. In order to ensure the carbon was fully incorporated and uniformly distributed within this sample,it was reground in acetone,pressed into a pellet,and sin-tered for an addition48hours at1200o C.The subsequent sample showed a further increase in the(110)peak position(Figure1)which yielded an inferred carbon content of x i2=0.069.This sample also showed decrease in T c(Figure 2),which is consistent with more carbon being incorporated in the structure. In addition to shifting to higher2θ,the(110)peak became sharper.The full-width-at-half-maximum(FWHM)decreased from0.221o to0.157o after the second sintering step.Therefore the second sintering step not only incorpo-rated more carbon but it appears to have resulted in a more uniform carbon distribution.It should be noted that after the two sintering steps the FWHM values of the MgB2peaks were comparable to those of the Si standard,in-dicating we have achieved a fairly high level of homogeneity.It is also worth noting that whereas the(110)peak shifts and sharpens as a result of a second reaction step,the(002)peak does neither,having FWHM values comparable to,but slightly larger than,the neighboring Si(311)peak after both reaction steps.This indicates that the c-axis spacing and periodicity are particularly insensitive to this degree of carbon doping and/or disorder.
Whereas the x i2value exceeds the nominal value of x n=0.05in the starting material,the presence of MgB4,as evidenced by strong peaks in the x-ray spectrum(Figure3)may account for the discrepancy if we assume no carbon enters the https://www.wendangku.net/doc/0c10213347.html,parison of the x-ray spectra for the single and two step reactions shows an increase in the intensity of the MgB4peaks after the second reaction step.This step was done without the addition of any extra Mg to compensate for potential Mg loss.It is possible that while sintering at1200o C,some Mg is driven out of the MgB2structure and this loss
results in conversion of MgB2to MgB4,with the excess Mg forming MgO and possibly condensing on the walls of the tantalum reaction vessel during the quench.To determine whether or not Mg loss is responsible for the apparent increase in the carbon content,the second sintering step for a sample with nominal concentration x n=0.05was carried out in an atmosphere of excess Mg vapor.This sample exhibited a T c of29.8K and a shift in the(110)x-ray peak yielding an inferred carbon concentration of x i2=0.050(Figure4), consistent with the nominal concentration.Thus the apparent di?erence in carbon content for samples which undergo a second sintering step without the presence of excess Mg to compensate for Mg loss relative to those which undergo the second sintering step with excess Mg is presumably the result of a?xed amount of carbon being incorporated into a decreased amount of MgB2.To avoid the potential creation of percolation networks of Mg within the samples we chose to perform the second sintering step without any additional Mg.
Using0.995purity boron and0.994purity B4C,an entire series with nominal carbon levels of x n=0,0.0125,0.025,0.035,0.05,and0.075was prepared using the two step reaction pro?le.The(002)and(110)x-ray peak positions for the entire series are plotted in?gure5.The(002)peak position is roughly constant for all carbon levels,consistent with the results found by Wilke et al.[7]and Avdeev et al.[20],which showed only a slight expansion along the c-axis for carbon doping levels up to10±2%.The(110)peak position shifts towards higher2θvalues as x is increased up to x n=0.05,at which point it appears to be https://www.wendangku.net/doc/0c10213347.html,ing the(110)peak position for the nominally pure sample as our standard,the inferred carbon concentrations for the entire series are x i2=0,0.01,0.034,0.044,0.069,and0.067.The samples with carbon concentrations saturating near x i2～0.07show an increase in the MgB2C2phase as a function of the nominal carbon content(Figure5b).Thus the excess carbon is precipitating out as MgB2C2.
Normalized magnetization measurements(Figure6)con?rm the highest two doping levels have incorporated roughly the same amount of carbon.Their transition temperatures all lie slightly below28K.De?ning T c using a2% screening criteria the x i2=0.069and0.067levels have T c values of27.5K and 27.8K respectively.For these higher doping levels,the nominal concentrations did not yield systematic increases in carbon level,but the change in the a-lattice parameter and T c are consistent with one another;i.e.samples which apparently incorporated more carbon had smaller a-lattice parameters and lower T c values.
This saturation near x i2～0.07is not entirely unexpected given the results of10±2%carbon incorporation using B4C reported by Ribeiro et al.[19] and Avdeev et al.[20].In optimizing the reaction,Ribeiro found that under certain conditions,T c values below the near22K reported for the optimal
24hours at1100o C reaction could be attained,suggesting higher carbon content phases may be metastable.To test whether the saturation we observed was an e?ect due to the use of a two step reaction,as opposed to the single step employed by Ribeiro et al.,we repeated their work,making a sample of nominal concentration Mg(B0.8C0.2)2using only Mg plus B4C.This sample underwent an initial48hour reaction at1200o C to form the superconducting phase and a second sintering for48hours at1200o C.After the?rst48hours at1200o C we?nd a superconducting phase with T c near22K,and a lattice parameter shift which yielded inferred values of x i1=0.092slightly less than the x i=0.10obtained by Ribeiro and coworkers using isotopically enriched11B4C as the carbon source[20].After the second sintering step,T c rises to27.9 K.The(110)x-ray peaks shifted to lower2θ,yielding an inferred x i2=0.065 (Figure7).Although two reaction steps were used,no change in the FWHM of the(110)peak was observed.As in the case of the x n=0.075sample,the decrease in carbon content could be due to carbon precipitating out in the form of MgB2C2.The relative intensity of the most prominent MgB2C2peak to that of MgB2approximately doubles going from the single step reaction to the two step reaction.In order to check if more carbon would be precipitated out in the form of MgB2C2by simply adding more sintering steps to the growth process an additional sample underwent a three step reaction:an initial48 hours at1200o C to form the superconducting phase followed by two additional sintering steps of48hours at1200o C.After this third reaction step,the sample exhibited a T c of27.9K and the(110)peak position yielded an inferred carbon content of x i3=0.064(Figure7).These values are comparable to our previous results with only two sintering steps.Thus the carbon content appears to saturate in the vicinity of an inferred carbon content of x i=0.065.The fact that saturation near x i2=0.065occurred for samples which had x i1both above and below this level indicates that in equilibrium the solubility limit for1200 o C reactions near1atm is in the range0.065 Transport measurements were made in order to determine the upper critical ?eld.An onset criteria was used in both resistance versus temperature and re-sistance versus?eld.Figure8a plots resistance versus temperature in applied ?elds up to14T and?gure8b plots resistance versus?eld at temperatures down to1.4K for the sample with a carbon level of x i2=0.069(x n=0.05).The zero?eld resistive transition has a width of less than1K,but this signi?cantly broadens as the strength of the applied?eld increases.The R vs.H measure-ments show a related broadening.For example,the width of the transition at 1.4K is nearly20T wide.This should be compared to the10T wide,ap-proximately linear transitions reported for5.2%carbon doped?laments[12]. Optical images taken under polarized light show the superconducting grains in the powder sample are5-10μm in size.In the case of the wire sample a majority of the grains are in the1-5μm range[12].We therefore ascribe the increased width of this transition to a combination of poor?ux pinning due to the large grain size associated with the high reaction temperature as well as to possible remaining inhomogeneities in carbon incorporation within the sample. H c2curves for x i2=0.034and x i2=0.069along with a pure wire[2]and a carbon doped wire with an inferred carbon content of x i=0.052[7]are plotted in?gure 9.The powder sample with x i2=0.034has a T c slightly less than that of the carbon doped wire with x i=0.052and an H c2(T=0)more than5T lower.This marked di?erence shows that for carbon doped samples made with di?ering nominal boron purities,T c alone is not a good caliper of H c2(T=0).The sample with inferred carbon content of x i2=0.069has a T c nearly7K below the aforementioned wire and an H c2(T=0)just https://www.wendangku.net/doc/0c10213347.html,paring the two powder samples to one another,we see an increase in the slope of H c2near T c for the higher doping level,which results in a higher H c2(T=0),consistent with our earlier?ndings[7,12]. Carbon has been shown to suppress T c at a rate of roughly1K/%C for up to 5%carbon substitution[12].The magnetization and transport measurements indicate T c of the powder samples made with the0.995purity Alpha Aesar boron is also being suppressed at a rate of roughly1K/%C,but relative to the suppressed,near37K,transition temperature of the nominally pure sample. The suppressed transition temperature of the nominally pure MgB2sample lies approximately2K below results obtained using high purity natural boron wires[23].MgB2made from lower purity boron has been shown to have lower transition temperatures[22].In?gure10,a comparison of T c versus|?a| for carbon doped samples prepared with lower purity boron to carbon doped wires made with high purity boron shows the manifold associated with the impure boron powder is shifted downward by approximately2K for all carbon levels.To con?rm that this di?erence is a result of the purity of the starting boron,a second set of two step process samples made with isotopically enriched 11B were measured.The results are included in?gure10.Also included is a set of carbon doped powders made by a plasma spray process[24].The agreement between the CVD wires,plasma spray powders,and11B samples shows that high purity boron in a variety of forms responds to carbon doping in a similar manner.These data also seem to indicate that there is some additional impurity in the0.995pure Alpha Aesar boron that is systematically suppressing T c. Figure11plots a comparison of the(002)and(110)x-ray peaks of pure MgB2using the three di?erent purity levels of boron.The sample made with the nominal0.995purity boron shows a shift of the(110)peak to higher2θby0.09o,which,if it were to be associated with carbon doping,would be consistent with carbon doping of approximately1.8%.This level far exceeds the stated carbon impurity level of0.25%in the0.995purity B as claimed in the certi?cate of analysis provided by Alpha Aesar.To check whether by using lower purity boron we have inadvertently doped with carbon to such a high level,we measured the resistive onset of superconductivity in an externally applied14T?eld for the nominal pure MgB2using the0.995purity boron and compared the temperature with those attained for carbon doped?bers reported in reference[12].MgB2?bers reacted at1200o C for48hours showed an onset of superconductivity in an externally applied magnetic?eld of14T at10.2K,14.8K,and18.5K for pure,0.6%,and2.1%carbon doping[12].If the shift of the(110)peak in the nominally pure MgB2made from0.995purity boron were a result of inadvertent carbon doping,we would expect an onset of superconductivity in an applied14T?eld at a temperature between15K and 18K.However,such a measurement yielded an onset near13K indicating if carbon is present as an impurity,it is less than0.6%,which consistent with the estimate provided by Alpha Aesar.Therefore the manifold of T c versus|?a| for the lower purity boron is shifted downward by some as of yet unidenti?ed impurity associated with the Alpha Aesar boron. H c2values were determined using an onset criteria in resistivity versus tem-perature and resistance versus?eld measurements.Pellets made using the isotopically enriched11B lacked structural integrity and were unsuitable for transport measurements.Therefore we could only attain H c2(T=0)values only for samples made from the CVD wires,plasma spray powders,and the0.995 purity boron(Figure12).For the dirtier,0.995purity powder,at a doping level of x i2=0.034the upper critical?eld agrees with the results of a carbon doped wire with an inferred carbon content of x i=0.038from reference[7].At doping levels near x i2=0.065,H c2values fall several Tesla below the manifold for”clean”carbon doped samples.In carbon doped MgB2,enhancement of H c2 due to scattering e?ects[11,26]competes with the suppression of T c caused by electron doping[11].By further suppressing T c by introducing additional impurities in the system,we may have limited the maximum H c2attainable through carbon doping. 4Conclusions We have established a method for synthesizing Mg(B1?x C x)2using a mixture of distilled magnesium,boron and the binary compound B4C.Impurities in the starting boron e?ect T c and the magnitude of the a-lattice parameter. By tracking|?a|and T c we were able to show that di?erent boron purities lead to di?ering T c(|?a|)manifolds.There appears to be a solubility limit in the carbon content for samples synthesized at1200o C and1atm near x～0.07.Lower purity boron in the starting material results in lower transition temperatures and appears to limit the maximum achievable upper critical ?eld. References [1]J.Nagamatsu,N.Nakagawa,T.Muranaka,Z.Takahiro,Y.Zenitani,and J. Akimitsu,”Superconductivity at39K in Magnesium Diboride”Nature(London, United Kingdom)410(2001)6824. [2]S.L.Bud’ko,C.Petrovic,https://www.wendangku.net/doc/0c10213347.html,pertot,C.E.Cunningham,P.C.Can?eld,M-H. Jung,and https://www.wendangku.net/doc/0c10213347.html,cerda,Phys.Rev.B63(2001)220503. [3]S.L.Bud’ko,V.G.Kogan,and P.C.Can?eld,Phys.Rev.B64(2001)180506. [4] A.K.Pradham,Z.X.Shi,M.Tokunaga,T.Tamegai,Y.Takano,K.Togano,H. Kito,and H.Ihara,Phys.Rev.B64(2001)212509. [5] A.V.Sologubenko,J.Jun,S.M Kazakov,J.Karpinski,and H.R.Ott,Phys. Rev.B65(2002)180505. [6]M.Eisterer,M.Zehetmayer,and H.W.Weber,Phys.Rev.Lett.90(2003) 247002. [7]R.H.T.Wilke,S.L.Bud’ko,P.C.Can?eld, D.K.Finnemore,Raymond J. Suplinskas,and S.T.Hannahs,Phys.Rev.Lett.91(2004)217003. [8]S.M.Kazakov,R.Puzniak,K.Rogacki,A.V.Mironov,N.D.Zhigadlo,J.Jun, Ch.Soltmann,B.Batlogg,and J.Karpinski,Phys.Rev.B71(2005)024533. [9]T.Masui,S.Lee,A.Yamamoto,K.Kajita,and S.Tajima,Physica C,Physica C412-414(2004)303-306. [10]R.Puzniak,M.Angst,A.Szewczyk,J.Jun,S.M.Kazakov,and J.Karpinski, cond-mat/0404579. [11]M.Angst,S.L.Bud’ko,R.H.T.Wilke,and P.C.Can?eld,Phys.Rev.B71(2005) 144512. [12]R.H.T.Wilke,S.L.Bud’ko,P.C.Can?eld, D.K.Finnemore,Raymond J. Suplinskas,and S.T.Hannahs,accepted for publication in Physica C. [13]A.Serquis,L.Civale,X.Z.Liao,J.Y.Coulter,Y.T.Zhu,M.Jaime,D.E. Peterson,F.M.Mueller,V.F.Nesterenko,and Y.Gu,Appl.Phys.Lett.82 (2003)2847. [14]P.Lezza,V.Ab¨a cherli,N.Clayton,C.Senatore,D.Uglietti,H.L.Suo,and R. Fl¨u kiger,Physica C401(2004)305. [15]X.L.Wang,Q.W.Yao,J.Horvat,M.J.Qin,and S.X.Dou,Supercond.Sci. Technol.17(2004)L21. [16]J.Wang,Y.Bugoslavsky,A.Ferenov,L.Cowey,A.D.Caplin,L.F.Cohen,J.L. McManusDriscoll,L.D.Cooley,X.Song,and https://www.wendangku.net/doc/0c10213347.html,rbalestier,Appl.Phys. Lett.81(2002)2026 [17]Y.Zhao,Y.Feng,C.H.Cheng,L.Zhou,Y.Wu,T.Machi,Y.Fudamoto,N. Koshizuka,and M.Murakami,Appl.Phys.Lett.79(2001)1154. [18]X.F.Rui,Y.Zhao,Y.Y.Xu,L.Zhang,S.F.Sun,Y.Z.Wang,and H.Zhang, Supercond.Sci.Technol.17(2004)689. [19]R.A.Ribeiro,S.L.Bud’ko,C.Petrovic,and P.C.Can?eld,Physica C384(2003) 227. [20]M.Avdeev,J.D.Jorgensen,R.A.Ribeiro,S.L.Bud’ko,P.C.Can?eld,Physica C387(2003)301. [21]S.Lee,Takahiko Masui,Ayako Yamamoto,Hiroshi Uchiyama,Setsuko Tajima, Physica C397(2003)7. [22]R.A.Ribeiro,S.L.Bud’ko,C.Petrovic,and P.C.Can?eld,Physica C382(2002) 194. [23]P.C.Can?eld,D.K.Finnemore,S.L.Bud’ko,J.E.Ostensen,https://www.wendangku.net/doc/0c10213347.html,pertot,C.E. Cunningham,and C.Petrovic,Phys.Rev.Lett.86(2001)2423. [24]J.V.Marzik,R.J.Suplinskas,R.H.T.Wilke,P.C.Can?eld,D.K.Finnemore,M. Rind?eisch,J.Margolies,and S.T.Hannahs,Physica C423(2005)83. [25]Z.Hoˇlanov′a,J.Kaˇc marˇc′ik,P.Szab′o,P.Samuely,I.Sheikin,R.A.Ribeiro,S. L.Bud’ko and P.C.Can?eld,Physica C404(2004)195. [26]A.Gurevich,Phys.Rev.B67(2003)184515. 51.051.552.052.559.059.560.060.561.0 I /I 0 2 Fig. 1.Powder x-ray di?raction (002)and (110)peaks for pure and nominal Mg(B .95C .05)2samples made using the 0.995purity boron as the starting mate- rial.The pure sample was reacted using two steps.For the carbon doped sample, the second sintering step shifts the (110)peak position to higher 2θindicating the incorporation of a higher carbon concentration. 051015202530354045 M /|M (5K )| Temperature (K) Fig.2.Normalized magnetization curves for nominal Mg(B .95C .05)2samples.The second sintering step lowers T c ,consistent with the incorporation of a higher carbon concentration. 3035404550556065 I /I 0 2 Fig.3.Normalized powder x-ray pattern for a sample with x n =0.05synthesized using a one step and a two step reaction.The two step sample clearly contains enhanced amounts of MgB 4. 051015202530354045 M /|M (5K )| Temperature (K) 51.051.552.052.559.059.560.060.561.0 (b) I /I 0 2 Fig.4.(a)Normalized magnetization curves and (b)x-ray (002)and (110)peaks for a sample of x n =0.05reacted using a two step process.If the second sintering step is performed without any excess Mg to compensate for potential losses,the resultant carbon content within the MgB 2phase is increased. 51.051.552.059.560.060.561.0 I /I 02 44.044.545.045.546.046.547.047.548.0 I /I 0 2 Fig.5.(a)Evolution of the (002)and (110)x-ray peaks for Mg(B 1?x C x )2samples with nominal x n =0,0.0125,0.025,0.035,0.05,0.075,and 0.20synthesized using a two step reaction.The shift of the (110)peak relative to that of the un-doped yields inferred carbon concentrations of x i 2=0.01,0.034,0.044,0.069,0.067,0.065.(b)For samples saturating near x i 2=0.07the excess carbon precipitates out in the form of MgB 2C 2as can be seen by the emergence of the MgB 2C 2(042)peak as a function of nominal carbon content. 2025303540 M /|M (5K )|Temperature (K) Fig. 6.Normalized magnetic transitions for the series of Mg(B 1?x C x )2with x i 2=0.01,0.034,0.044,0.069,0.067,0.065,synthesized with 0.995purity B and reacted using a two step process. 51.051.552.052.559.560.060.561.0 I /I 0 2 15202530 M /M |(5K )|Temperature (K) Fig.7.(a)(002)and (110)x-ray peaks for nominal Mg(B 0.8C 0.2)2using B 4C as the boron and carbon source and reacted using 1,2,and 3step reaction processes.(b)Normalized magnetic transitions for these samples. 0510152025303540 (a) R e s i s t a n c e ()Temperature (K)