Chemical Engineering Journal 256 (2014) 32–38

Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier .com/locate /ce j

Nano-sized boron synthesis process towards the large scale production

http://dx.doi.org/10.1016/j.cej.2014.06.1181385-8947/� 2014 Elsevier B.V. All rights reserved.

⇑ Corresponding author. Tel.: +39 010 6598790; fax: +39 010 6598732.E-mail address: maurizio.vignolo@spin.cnr.it (M. Vignolo).

Maurizio Vignolo a,⇑, Gianmarco Bovone b, Davide Matera a, Davide Nardelli c,Cristina Bernini a, Antonio Sergio Siri a,b

a CNR-SPIN, C.so Perrone 24, 16152 Genova, Italyb Physic Department of Genoa University, Via Dodecaneso 33, 16146 Genova, Italyc Columbus Superconductors, Via delle Terre Rosse 30, 16133 Genova, Italy

h i g h l i g h t s

� An innovative process to synthesize boron powder has been reported.� The described process has demonstrated its suitability for the large scale application.� Using the new process, the MgB2 powder can be improved through a homogenous dispersion of dopants.

a r t i c l e i n f o

Article history:Received 14 May 2014Received in revised form 26 June 2014Accepted 30 June 2014Available online 9 July 2014

Keywords:Nano-sized boronSynthesis processMgB2

a b s t r a c t

In the present paper a new process for large scale production of nano-sized boron is reported. The processcan be summarized in several steps: boron oxide solubilization in hot water, cryogenic freezing of liquidphase, freezing–drying process, magnesiothermic reduction of boron oxide, boron purification. Each stepis described in order to show the innovations and then the purified boron has been employed to synthe-size the superconducting MgB2 powder. It is worth to note that for the first time the same MgB2 precur-sors were used to prepare the superconducting phase following four different techniques and the resultsdirectly compared. So several MgB2 conductors were prepared applying different techniques, ex-situ,in-situ, via MgB4 and RLI, and then their superconducting properties investigated. Furthermore morphol-ogy, grain size and purity of B and MgB2 powder were analyzed by SEM analysis and X-ray diffractiontechnique.

� 2014 Elsevier B.V. All rights reserved.

1. Introduction

Boron is a relatively rare element in the Earth’s crust, represent-ing only 0.001% of the crust mass and it does not appear on earth inelemental form but is found combined with others elements inborax, boric acid, colemanite, kernite, ulexite and borates. Despitethe rapid growth of boron applications, limited progresses havebeen made in the synthesis of nanometric boron powder, espe-cially for doped powder. The most important method to produceelemental boron involved reduction of boron oxide (B2O3) withmetals, such as Mg or Al, reaching a purity grade between 90%and 98%. Pure boron can be prepared by reducing volatile boronhalides with hydrogen at high temperatures. Ultrapure boron foruse in the semiconductor industry is produced by the decomposi-tion of diborane at high temperatures and then further purifiedwith the zone melting. Nearly all synthetic methods for boron

powder are based on the technologies that involve the usage ofgas phase decomposition of highly toxic and flammable precursorygases (B2H6, H2, BCl3) under severe reaction conditions, and thenthey are inevitably expensive and dangerous processes. In orderto provide a broader view about the techniques of B powder synth-esis, the main synthesis reactions of B have been listed in thefollowing:

Reduction of boron compounds by:

– metallothermic reaction: i.e. B2O3(l) + 3 Mg(l) ? 3MgO(s) + 2B(s)

at 1000 �C;– electrolysis from a melt: B2O3–K2O–KF or B2O3–KBF4–KF at

800–900 �C;– hydrogen: i.e. BCl3(g) + 3/2 H2(g) ? 3HCl(g) + B(s) on hot W (or Ta)

filament at 1300 �C;

Thermal decomposition of boron compounds:

– boron hydrides: B2H6(g) ? 2B(s) + 3H2(g) on Ta filament at950 �C;

H3BO3 solubilizationwith deionized hot H2O(doping agent addition)

Cryogenic freezing in LN2

Freezing drying

Nano-structured B2O3 (doped)

mixing with Mg and reduction for 2 h at 1000 °C in Ar flow

MgO is removed by several leaching and rinsing (HCl 18%

(V/V) and hot water) thenheat treated in Ar/H2 flow at

1100°C

Nano-sizedboron (doped)

mixing with Mg and reacted in Ar flow at

920°C for 1 h

MgB2 (doped) powder for ex-situ P.I.T. tape

manufacturing

1 2 3

4

5

Fig. 1. Stages of the B powder synthesis and MgB2 preparation.

M. Vignolo et al. / Chemical Engineering Journal 256 (2014) 32–38 33

– boron halides: BX3(g) ? B(s) + 3/2 X2(g) on W (or Ta) filament at1400 �C.

In the present paper we propose a new inexpensive technologyfor carbon-doping (C-doping) and nano-sizing of boron based onfreezing-dry process step before the Moissan’s magnesiothermicreduction of boron oxide. We can speculate that freeze drying (lyo-philization) is a technique widely employed by pharmaceutical andfood industries to remove water by sublimation from frozen phase,so it results a relatively inexpensive process compared to theplasma synthesis process [1].

The innovative preliminary step considers the preparation of aB2O3 or boric acid (H3BO3) solution in water. At this step, the intro-duction of a soluble organic molecule (carbohydrate), or an insolu-ble inorganic compound (SiC) or element (C) can be done. In thisway it is possible to have a homogeneous solution or dispersionof the C-source in the H3BO3 medium at the end of the freeze-dry-ing process. The C-doping is particularly useful for high magneticfield applications of the MgB2 conductor [2–5]. In order to havethe most possible fine precursor powder, the freezing process ofthe liquid solution has to be leaded in a cryogenic liquid (liquidN2). Following the new boron synthesis process is possible toobtain a suitable precursor for MgB2-based conductor manufactur-ing with the same performance of conductors prepared using MgB2

powders synthesized from commercial B precursor. In particular inprevious articles we shown that the laboratory made B has thesame critical current density (Jc) performance of milled MgB2 pow-ders [6–8].

The grain size of MgB2, as well as the doping of MgB2, plays afundamental role in critical current performances. In fact, it is wellknown that MgB2 follows a grain boundary pinning mechanism,and then finer the powder higher the pinning force and betterin-field Jc behavior will be shown by the conductor. For the aboveconsiderations the grain size of B can play an important role,furthermore also the reactivity versus the magnesium will beimproved using nano-sized B and the MgB2 synthesis temperatureand reaction time can be lowered.

In the present paper we report the scaling-up of the laboratoryprocess, producing 0.1 kg both of pure and Si-doped B. It is worthto note that: 0.5 kg of Mg and 2.3 kg of H3BO3 are needed in orderto produce 0.1 kg of B. The resulting raw B must be processed withacid leaching to remove 2.7 kg of secondary products, MgO, B2O3,and MgB2. A big amount of mass is involved in the process heredescribed. Anyway, it is not a problem to process a bigger amountof reactants for a chemical factory, which could obtain 10 kg (ormore) per batch. 10 kg of B powder give 20 kg of MgB2 useful for20 km of cable.

It is interesting to note that the reduction with elemental Mg iscarried out with excess of B2O3 respect to the Mg, in order to limitthe side reaction, so instead to use the stoichiometric ratio 1:3(B2O3:Mg) we adopt 1:1.1 (B2O3:Mg). In this way the side reactiondue to direct reaction between Mg and B (just formed) is attenu-ated. However, the reaction has a low yield with respect to thestarting amount of reactants. In order to have 1 kg of B will benecessary to react 13 kg of B2O3 with 5 kg of Mg. The theoreticalproduct yield of the stoichiometric ratio using 5 kg of Mg is1.4827 kg of B; instead the yield of the process using the B2O3

excess is only 1 kg. The 67% of yield is principally due to the con-temporary formation of magnesium borate (xB2O3 * yMgO) andmagnesium boride (MgB2) of the respective side reactions.

Fig. 2. Pressure–temperature diagram of water.

2. Experimental details

The process to prepare boron can be summarized in the follow-ing steps:

(1) Solubilization of H3BO3 (as well as of the doping source),(2) Cryogenic freezing,(3) Freeze-drying,(4) Reduction to row elemental boron,(5) Acid leaching and heat treatment at high temperature of B.

The last step is the conversion of B (using Mg) into MgB2 phase,which is used for the MgB2 conductor manufacturing by ex-situpowder in tube (P.I.T.) process, or other techniques.

Steps (1), (2) and (3) represent the innovation introduced intothe Moissan’s process [9].

The process is summarized in Fig. 1.

2.1. B2O3 precursor preparation

The H3BO3 (2.3 kg) was dissolved in boiling deionized water(10 l) to maximize g/l ratio, i.e. 276 g/l at 100 �C against 46.5 g/lat room temperature. At this step some doping agents can beadded, for this work was chosen 10% w/w SiO2. So the processwas used to prepare two different nano-sized B powders: pure Band 10% w/w Si-doped B. Independently, each solution was cryo-genically frozen in liquefied nitrogen (LN2) to obtain a solid phaseuseful for freezing–drying process. A sloping plan has been used tokeep a polypropylene LN2 container in front of bain-marie contain-ing the boiling solution. Then the boiling solution is sprayeddirectly into the LN2 using compressed air (2 bars). Cryogenicfreezing has the double purpose to maintain the chemical homoge-neity during the solid phase formation and to lead the biggest sur-face area of solid phase useful for the sublimation process. Thesolid phase is then placed in the Coolsafe™ freeze-dryer (model55–4) in order to remove water by sublimation, experimental con-ditions are summarized in Fig. 2.

120 cm

Fig. 3. The beaker used for hot acid leaching.

(b) (a)

34 M. Vignolo et al. / Chemical Engineering Journal 256 (2014) 32–38

Sublimation process is necessary to avoid the liquid phase for-mation and keep the homogeneous dispersion of the doping agent,which will be not segregated from B2O3 during crystallization. Thelyophilization process is developed at room temperature and0.03 * 10�2 bar in pressure by a rotary pump, the process is main-tained under the triple point of water, see Fig. 2. Water vapor iscondensed in a refrigerated chamber at �55 �C. At the completionof the process, the so treated product will have retained its form,volume and original structure, as well as its physical and chemicalproperties. If necessary the product can be stored (packaging iseffective to the reduction of moisture migration) for an indefiniteperiod of time.

The main chemical reactions involved in step 2.1 are:

BðOHÞ3ðsÞ þH2OðlÞ �!

BOðOHÞ�2ðaqÞ þH3OþðaqÞ

ðKa1 ¼ 5:8� 10�10; pKa1 ¼ 9:24Þ ð1Þ

BOðOHÞ�2ðaqÞ þH2OðlÞ �!

BO2ðOHÞ2�ðaqÞ þH3OþðaqÞ

ðKa2 ¼ 4� 10�13; pKa2 ¼ 12:4Þ ð2Þ

BO2ðOHÞ2�ðaqÞ þH2OðlÞ �!

BO3�3ðaqÞ þH3OþðaqÞ

ðKa3 ¼ 4� 10�14; pKa3 ¼ 13:3Þ ð3Þ

And the phase transition of water during the freezing-dryprocess:

H2OðsÞ ! H2OðvÞ ð4Þ

200 cm 20 60

(c)

2.2. Synthesis of nano-sized boron from boron oxide precursor

The lyophilized pure and Si-doped B2O3 are reduced by Mg intwo different stainless steel crucibles for 2 h at 1000 �C under Arflow. Internally the crucible was upholstered with a Nb sheet inorder to minimize the reaction of the reagents with the stainlesssteel. The reduction reaction is leaded in excess of B2O3 to mini-mize the magnesium borides formation following the molar ratio1:1.1 (B2O3:Mg) instead of the stoichiometric ratio 1:3 (B2O3:Mg).

The main chemical reactions involved in step 2.2 are:

3MgðlÞ þ B2O3ðlÞ �T

3MgOðsÞ þ 2BðsÞ ðmain reactionÞ ð5Þ

MgðlÞ þ 2BðsÞ �T

MgB2ðsÞ ðside reactionÞ ð6Þ

xMgOðsÞ þ yB2O3ðsÞ �T ðxMgO � yB2O3ÞðsÞ ðside reactionÞ ð7Þ

Fig. 4. The filtering system. (a) Assembled, (b) top view of the Buchner with paperfilter, (c) filter holder.

2.3. Boron purification

Each kind of raw B was acid leached in hydro-chloridric acid(18%) a 90 �C in home-made polypropylene beaker. The polypropy-lene beaker is sealed at both ends by a flange system. The uppercap has an inlet/outlet valve to permit charge and discharge ofsolutions and a security valve to keep a constant internal pressure.The beaker was placed into a home-made bain-marie, an electricalheater controlled by a thermocouple keeps the temperature at90 �C. The volumetric capacity of the beaker is 12 l, Fig. 3. After5 h of heating process, the solution is transferred by a peristalticpump into home-made filtering system. The filtering system is apolypropylene cylinder (2 m in height and i.d. of 0.25 m (Fig. 4a)with a Buchner’s filter at the bottom side (Fig. 4a and b). The filteris a Whateman Filter Paper No. 42, 0.24 m in diameter) fixed by afilter holder (Fig. 4c). At the end of the filtration the raw B is rinsedwith deionized hot water in order to remove MgCl2. This operationis repeated for 4 times, and then a cycle is repeated using

hydro-fluoric acid (20%) at room temperature. A final leaching pro-cess in hydro-chloric acid is performed before to apply the lastrinse process up to reach neutral pH. The followed purification pro-cess is in agreement with Moissan’s procedure.

The main chemical reactions involved in step 2.3 are:

MgOðsÞ þ2H3OþðaqÞ þ2X�ðaqÞ ! Mg2þðaqÞ þ2X�ðaqÞ þ3H2OðlÞ ðX¼ Cl; FÞ

ð8Þ

MgB2ðsÞ þ 2H3OþðaqÞ þ 2X�ðaqÞ þ 4H2O

! Mg2þðaqÞ þ 2X�ðaqÞ þ 2H3BO3ðaqÞ þ 4H2ðgÞ ð9Þ

For Eqs. (8) and (9) formally the chlorine ion on the left andright side of the yield sign should be elided.

Table 1Description of the samples.

Sample Method MgB2 Deformation Synthesis/sintering [�C–h]

1 Ex-situ Pure Groove rolled 900–1/920–0.32 In-situ Pure Groove rolled 800–13 1/2–1/2 Pure Groove rolled 800–14 RLI Pure Groove rolled 700–15 Ex-situ Si-doped Groove rolled 900–1/920–0.3

M. Vignolo et al. / Chemical Engineering Journal 256 (2014) 32–38 35

ðxMgO � yB2O3ÞðsÞ þ 2xH3OþðaqÞ þ 3yH2O

�! xMg2þðaqÞ þ 2yH3BO3ðaqÞ þ 3xH2O ð10Þ

The formation of H3BO3 is due to the presence of HCl, and in thesubsequent washing with deionized water it is removed followingEq. (1).

2.4. Heat treatment at high temperature

The purified B is heat treated at high temperature (1100 �C) for10 h under Ar flow [7] in a furnace directly connected with a glove-box, at the end of the process the powder can be stored under Aratmosphere or directly used to prepare MgB2 powder. This precau-tion avoids the oxidization.

2.5. MgB2 synthesis

Four different techniques were used to prepare the MgB2-basedconductors.

2.5.1. Ex-situ MgB2

Nano-sized B, pure and Si-doped were mixed in stoichiometricratio with elemental Mg and reacted at 900 �C for 1 h under Ar flowin a stainless steel crucible, following the standard proceduredeveloped in our laboratory [10]. The crucible was upholsteredwith a Nb sheet. The synthesis was carried out in the furnace con-nected to the glove-box.

2.5.2. In-situ MgB2

The reagents products were mixed in stoichiometric ratio andused to fill the Ni sheath. The synthesis of MgB2 was done afterthe deformation process at 800 �C for 1 h in Ar flow [11].

2.5.3. MgB2 via MgB4 or 1/2–1/2 methodThe procedure was already described in reference [12]. Synth-

esis of MgB2 was done after the deformation process at 800 �Cfor 1 h under Ar flow. This method is a hybrid method, 1/2 ex-situ(MgB4 was synthesized with ex-situ approach) and 1/2 in-situ(MgB2 with in-situ method).

2.5.4. Reaction liquid infiltration (RLI)Following Giunchi’s methodology [13] nano-sized B was

pressed in the Ni tube (with internal Fe barrier) then a hole wasdone in order to insert a Mg rod. So the MgB2 phase wassynthesized at the end of the cold working process at 700 �C for1 h in Ar flow.

Each typology of MgB2 phase was reacted at the best corre-sponding temperature and time condition. The Si-doped samplewas prepared only for the ex-situ method.

Fig. 5. SEM image at 50,000� of the B

2.6. Wire manufacturing

A Ni tube was employed as metal sheath (i.d. = 8 mm, e.d.12 mm) and filled by the MgB2 powder, both sides of the Ni sheathwere sealed with Sn cap to preserve the Ar atmosphere of theglove-box, then avoiding the oxygen contamination during thedeformation process. An inner Fe barrier was added in order tominimize the reaction between MgB2 and Ni. The outer diameterwas reduced to a wire 7000 mm in length with a square sectionof 1 mm � 1 mm by groove rolling machine. Finally short samples,12 cm in length were heat treated in Ar flow to synthesize (or tosinter) the MgB2 phase and recover the collected cold deformationstress. Ex-situ at 920 �C for 20 min, RLI at 700 �C for 60 min. and800 �C for 60 min. for the other samples. The samples preparedfor this paper are summarized in Table 1.

2.7. Samples characterization

Morphology and grain size of B and MgB2 powder were charac-terized by SEM microscopy. The powder samples were analyzed byXRD technique, using a PHILIPS diffractometer (Bragg–Brentanogeometry, Ni filtered Cu Ka radiation, 40 kV and 30 mA), in orderto check the purity phase. Short pieces of wires, about 6 mm inlength, were employed for magnetization vs. magnetic field mea-surements with a commercial 5.5 MPMS Quantum Design Squidmagnetometer, and Jc values were calculated using the Bean’smodel [14].

The four probe method was applied to measure the resistivitydrop on 3 cm long samples, in order to establish the Tc onset.

3. Results

Fig. 5 reports the SEM images at 50000x of the B powders, pureand Si-doped.

Both images show very fine B (pure and Si-doped) powderswith a homogeneous grain size distribution. Some agglomeratesof very fine particles are present.

Statistical analysis on SEM images confirms an average grainssize of 80 nm in agreement with analysis of previous paper [15].

powders: (a) pure, (b) Si-doped.

10 20 30 40 50 60 70 80 90

0

1000

2000

3000

4000

5000

+@

@@@

@

@

@

@

°° ° ° ° °

°

°#### **

***

**

*

*

Inte

nsi

ty [

a. u

.]

2θθ [degree]

Si-doped B

+ SiO2

Si-doped MgB2

* MgB2

° MgO# Mg@ Mg

2Si

*

°@

Fig. 8. Xrd pattern of Si-doped B and corresponding Si-doped MgB2 powder.

36 M. Vignolo et al. / Chemical Engineering Journal 256 (2014) 32–38

The morphological aspect of pure MgB2 is given in Fig. 6,Si-doped MgB2 image has been omitted because it has the samemorphology and grain size. Some big agglomerates of very fineMgB2 particles are present. It is worth to note that the ex-situMgB2 synthesis process (as well as for the other techniques) hasnot been optimized yet for the sub-micrometric B powder. Then,using the standard synthesis procedure [10] the very fine grain sizeof B (80 nm) has been lost and MgB2 particles size was increased 10times in size. The statistical analysis given an average grain size of600 nm, in good agreement with results reported in [15] both forpure and Si-doped ex-situ powder.

Fig. 7 shows the xrd patterns of pure B and the correspondingpure MgB2 powder. The xrd spectrum of B powder is the typicalone for an amorphous substance, no secondary phase was found.The MgB2 pattern shows the presence of very small amount ofunreacted Mg and some MgO. It is impossible to say if the MgOis already present inside B or the amorphous B undergoes oxidisa-tion process during the measurements.

Fig. 8 reports the xrd patterns of Si-doped B and the correspond-ing Si-doped MgB2. The intensity of MgB2 peaks is smaller than thepure MgB2, in fact some Mg reacts with Si to give Mg2Si instead togive MgB2. The presence of Si is confirmed in B pattern, where thepeak of SiO2 is present at 26.3 2h degree. The presence of a biggeramount of MgO than the pure can be explained taking into accountthat the SiO2 reacts with Mg giving Mg2Si and MgO.

The longitudinal sections of the wires, acquired by SEM, arereported in Fig. 9. Some metallic inclusions (brightest particles inFig. 9) are due to the metallographic preparation and visible bothin the MgB2 core and in the resin matrix. The best section corre-sponds to the ex-situ technique; in fact it shows the most

Fig. 6. SEM image at 20,000� of the pure MgB2.

10 20 30 40 50 60 70 80 90

0

1000

2000

3000

4000

5000

6000

7000

8000

°

°##°

#°

* ** **

**

*

*

Inte

nsi

ty [

a. u

.]

2θ [degree]

pure MgB2

pure B

* MgB2

° MgO# Mg

*

Fig. 7. The xrd pattern of pure B and corresponding pure MgB2 powder.

compacted MgB2 phase. In particular the RLI procedure leads tosome big voids (black holes). Instead the voids for the 1/2–1/2method, as well as for the in-situ, are present in ribbon shape.

Fig. 10 reports the magnetic Jc measurements at 5 K of thewires. The best behavior is shown by the RLI conductor in spiteof the worse section, reaching the best Jc value of 105 A/cm2 at4 T, one order in magnitude bigger than the other samples(104 A/cm2). Instead the others techniques have more or less thesame behavior. The in-situ process shows the worse Jc performanceat low field.

In Fig. 11 is reported the comparison of Jc behavior between thepure and Si-doped ex-situ MgB2 conductors. The doped sampleshows a Jc value 4 times bigger than the un-doped at 4 T. Bothsamples have a Jc value of 3.1 * 105 A/cm2 at 1 T.

The resistivity vs. temperature measurements, by four probemethod, are reported in Fig. 12. The best critical temperature (Tc)onset (39.5 K) is shown by the 1/2–1/2 sample, Si-doped andun-doped ex-situ samples have the same Tc value and transitionshape. The lowest Tc onset is shown by samples in-situ and RLI.

4. Discussion

4.1. Undoped wire samples

For the first time four different techniques were used to prepareMgB2-based conductors using the same B precursor in order to pro-duce the superconducting phase. From the comparison of the Jc

values and Tc data emerges clear that the best Jc value correspondsto the RLI conductor, which has the worst Tc value and awful crosssection. Probably this Jc behavior is due to the highest connectivityreached rather that to the quality of the MgB2 phase. In fact it canoccur that following the RLI technique an intrinsic phase gradientis done. Before MgB2 synthesis a Mg rod is surrounded byB powder. After the reaction, the MgB2 occupies the volume occu-pied by B powder before the reaction, whereas a hole substitutesthe Mg volume. The Mg diffusion into B leads a Mg gradient intothe MgB2 phase. Near the void left by Mg, MgB2 phase is richerin Mg, instead distant by it MgB2 phase is richest in B (or poorestin Mg) [13]. Anyway, the 1/2–1/2 method has a very high Tc onsetvalue (39.5 K) and its behavior is better than both the ex-situ andin-situ sample. A so high Tc onset value indicates that the B precur-sor is really good, but the different production techniques play adifferent role on the MgB2 phase formation. The worst Jc

correspond to the in-situ wire, which has the worst density anda bad Tc onset value.

We must take into account that the in-situ process is affected bya low mean density respect to the ex-situ method, the same is truecomparing the in-situ and ex-situ techniques to the 1/2–1/2

Fig. 9. Longitudinal section of the wires: (a) RLI, (b) 1/2–1/2, (c) in-situ and (d) ex-situ.

0 1 2 3 4 5103

104

105

106

J C [

A/c

m2 ]

μ0H [T]

RLI 1/2-1/2 in-situ ex-situ

Fig. 10. Magnetic Jc values for the MgB2 conductors: ex-situ, in-situ, 1/2–1/2 andRLI.

0 1 2 3 4 5103

104

105

106

J C [

A/c

m2 ]

μ H [T]

Si-doped MgB2

pure MgB2

Fig. 11. Magnetic Jc values of the ex-situ MgB2 wires: pure and Si-doped.

33 34 35 36 37 38 39 40 41

0.0

0.1

0.2

0.3

0.4

0.5R

esis

tivi

ty [

μΩ c

m]

T [K]

Pure Rli 1/2-1/2 in-situ ex-situ

Si-doped ex-situ

Fig. 12. Transition from normal to superconducting state of each sample.

M. Vignolo et al. / Chemical Engineering Journal 256 (2014) 32–38 37

method (as well as to the RLI). So the Jc and the Tc trends of thethree samples can be explained. Very good MgB2 phase compactionhas been reached in the RLI sample, leading to a very good Jc beha-vior. In this case it is evident that the connectivity of the sample

plays a more important role than the purity grade of the MgB2

phase.

4.2. Ex-situ wire: pure and Si-doped samples

From the resistivity vs. temperature measurements we can seethe same behavior for the pure and Si-doped sample, but the Jc per-formance is better than the undoped. This is probably due to ahomogeneous dispersion of defects, both in a proper amount thatdimensional, inside the MgB2 matrix. On the other hand the newsynthesis method for producing doped and nanostructured Bpowders works very well, making possible a development of thatprocess on large scale production.

4.3. Influence of the process variables and possible optimization

The above described process has two bottle-necks. The first isthe freeze-drier actually employed, which is suitable for laboratoryscale and limits the production rate on 1 month to remove 10 l ofwater. This is easily overcome by adopting an industrial freeze-drier able to remove 10–20 l of water in 24 h [16]. In fact, theindustrial freeze-driers have several shelves with a total surface

38 M. Vignolo et al. / Chemical Engineering Journal 256 (2014) 32–38

area of 2 m2, or more. Furthermore in order to increase the subli-mation process a radiant heat plate is placed above the top shelffor uniform drying on each shelf. With this improvement we fore-cast that the boron precursor useful for 100 g of elemental B can beprepared in 1–2 days, instead of the current 30 days. The secondbottle-neck is the filtration process. In fact we adopt a filtrationprocess carried out by gravity, where the raw B powder is deposedat the bottom of the filter lowering the filtration rate, Fig. 4. Theelapsed time for the process generally need of 2 weeks. In orderto overcome this issue we are taking into account a filter-presstechnology to decrease the separation time down to few hours. Fil-ter-press is able to process 500–800 l/h adopting paper filter(200 mm � 200 mm) and to retain particles with grain size of0.2 lm [17].

5. Conclusions

For the first time a direct comparison among different wiremanufacturing processes, using the same precursors (B and Mg)has been shown. The influence of densification on Jc performanceseems to have a more important role than the MgB2 phase purity.The new method to synthesize nanostructured and doped B hasdemonstrated to work very well and it shown the possibility toincrease the amount of product on large scale, passing from few[18] grams to 100 g.

Acknowledgments

We wish to thank Gianfederico Vivado owner of the Ratto OMSand Ivan Fontana for their financial and technical support rela-tively to the boron purification process.

References

[1] J.V. Marzik, R.C. Lewis, M.R. Nickles, D.K. Finnemore, J. Yue, M.J. Tomsic, M.Rindfleisch, Plasma synthesized boron nano-sized powder for MgB2 wires, AIPConference Proceedings, Adv. Cryogenic Eng. 1219 (2010) 295–301.

[2] S.X. Dou, S. Soltanian, J. Horvat, X.L. Wang, S.H. Zhou, M. Ionescu, H.K. Liu, P.Munroe, M. Tomsic, Enhancement of the critical current density and flux

pinning of MgB2 superconductor by nanoparticle SiC doping, Appl. Phys. Lett.81 (2002) 3419.

[3] J. Wang, Y. Bugoslavsky, A. Berenov, L. Cowey, A.D. Caplin, L.F. Cohen, J.L. MacManus-Driscoll, L.D. Cooley, X. Song, D.C. Larbalestier, High critical currentdensity and improved irreversibility field in bulk MgB2 made by a scalable,nanoparticle addition route, Appl. Phys. Lett. 81 (2002) 2026.

[4] M.D. Sumption, M. Bhatia, M. Rindfleisch, M. Tomsic, S. Soltanian, S.X. Dou,E.W. Collings, Large upper critical field and irreversibility field in MgB2 wireswith SiC additions, Appl. Phys. Lett. 86 (2005) 092507.

[5] Y. Ma, H. Kumakura, A. Matsumoto, H. Hatakeyama, K. Togano, Improvementof critical current density in Fe-sheathed MgB2 tapes by ZrSi2, ZrB2 and WSi2

doping, Supercond. Sci. Technol. 16 (2003) 852–856.[6] G. Bovone, M. Vignolo, C. Bernini, S. Kawale, A.S. Siri, An innovative technique

to synthesize C-doped MgB2 by using chitosan as the carbon source,Supercond. Sci. Technol. 27 (2014) 022001.

[7] M. Vignolo, G. Bovone, C. Bernini, A. Palenzona, S. Kawale, G. Romano, A.S. Siri,High temperature heat treatment on boron precursor and PIT processoptimization to improve the Jc performance of MgB2-based conductors,Supercond. Sci. Technol. 26 (2013) 105022.

[8] G. Romano, M. Vignolo, V. Braccini, A. Malagoli, C. Bernini, M. Tropeano, C.Fanciulli, M. Putti, C. Ferdeghini, High-energy ball milling and synthesistemperature study to improve superconducting properties of ex-situ tapes andwires, IEEE Trans. Appl. Supercond. 19 (2009) 2706–2709.

[9] R. Naslain, Boron syntheses, in: P. Hagenmuller (Ed.), Preparative Methods inSolid State Chemistry, Academic Press INC., New York, 2012, pp. 440–483.

[10] M. Vignolo, G. Romano, A. Malagoli, V. Braccini, M. Tropeano, E. Bellingeri, C.Fanciulli, C. Bernini, V. Honkimaki, M. Putti, C. Ferdeghini, Role of the grainoxidation in improving the in-field behavior of ex-situ tapes, IEEE Trans. Appl.Supercond. 19 (2009) 2718–2721.

[11] P. Lezza, C. Senatore, R. Flükiger, Improved critical current densities in B4Cdoped MgB2 based wires, Supercond. Sci. Technol. 19 (2006) 1030.

[12] D. Nardelli, D. Matera, M. Vignolo, G. Bovone, A. Palenzona, A.S. Siri, G. Grasso,Large critical current density in MgB2 wire using MgB4 as precursor,Supercond. Sci. Technol. 26 (2013) 075010.

[13] L. Saglietti, C. Orecchia, Alessandro F. Albisetti, E. Perini, G. Giunchi,Microstructure of the MgB2 wires resulting by the infiltration process, IEEETrans. Appl. Supercond. 21 (2011) 2655–2658.

[14] C.P. Bean, Magnetization of hard superconductors, Phys. Rev. Lett. 8 (1962)250–253.

[15] M. Vignolo, G. Bovone, E. Bellingeri, C. Bernini, G. Romano, M.T. Buscaglia, V.Buscaglia, A.S. Siri, Size determination of superconducting MgB2 powder frommagnetization curve, image analysis and surface area measurements,Supercond. Sci. Technol. 27 (2014) 065007.

[16] <http://www.pharma-tech.it/magnum-series-freeze-dryers>.[17] <http://www.akciza.net/en/plate-filters-press-filters/2030-colombo-18-

automatic-press-filter.html#.U6g9XEDap_Z>.[18] M. Vignolo, G. Romano, A. Martinelli, C. Bernini, A.S. Siri, A novel process to

produce amorphous nanosized boron useful for synthesis, IEEE Trans. Appl.Supercond. 22 (2012) 6200606.