JOURNAL OF APPLIED ELECTROCHEMISTRY 28 (1998) 1147ą1151
TECHNICAL NOTE
Electrosynthesis attempts of tetrahydridoborates
E. L. GYENGE, C. W. OLOMAN
Department of Chemical Engineering, University of British Columbia, Vancouver, BC, V6T 1Z4, Canada
Received 1 August 1997; accepted in revised form 12 December 1997
Keywords: Tetrahydridoborates, electrosynthesis
1. Introduction chemical route as compared with the chemical
synthesis, there is little information in the open lit-
Tetrahydridoborates (i.e., commonly but less accu- erature regarding the electroreduction of borates to
rately called borohydrides, BH 4 ) are versatile re- borohydrides. Generally speaking the electrochemis-
ducing agents in various organic and inorganic try of boron compounds is largely based on electro-
processes [1]. The most important manufacturing oxidations [7]. However, in a paper devoted to the
technology of NaBH4 is based on the reaction of voltammetric determination of BH 4 , Mirkin and
trimethyl borate B(OCH3ą3, with sodium hydride at Bard brieŻy mentioned the complete absence of
about 250 C [2]. Electrosynthesis has been examined borohydride during the electroreduction of sodium
as a potentially simpler process for production of metaborate [8].
NaBH4 and a number of patents were granted in the The aim of the present study was to verify the
period 1958ą1990 [3ą6]. The 1958 patent by Hu� and above patents and to ascertain the possibility of
Adams [3] is rather an electrochemical metathesis borohydride electrosynthesis under diverse experi-
reaction where the sodium from sodium borohydride mental conditions.
is replacesd in nonaqueous media (e.g., methylamine)
by another metal (e.g., Mg or Ca) which represents
2. Experimental apparatus and procedures
the sacri�cial anode of the electrochemical cell.
However, the rest of the patents [4ą6] claim the
2.1. Analysis of borohydrides
possibility of the electrochemical reduction of both
alkali metal [4, 6] and organic borates [5, 6] to the
In order to avoid erroneous results leading to false
corresponding borohydrides.
conclusions, each sample was analysed by two or
The patent by Cooper [4] suggests an aqueous
three di�erent methods. Moreover, samples from
catholyte composed of at least 1% by weight sodium
blank experiments (i.e., either without current or
or potassium metaborate BO 2 ą. According to the
borate) were taken and analysed to �lter out possible
patent [4] the cathode material should be either an
interferences in the analytical procedure. The fol-
e�ective hydrogenation catalyst (e.g., nickel, nickel
lowing methods of borohydride analysis were em-
boride, Raney nickel, platinum, cobalt, cobalt boride)
ployed:
or mercury. The recommended anolyte was sodium
hydroxide which was separated from the cathode (i) The iodate method [9], which is based on the
compartment by a cation exchange membrane. By reaction of BH 4 with IO 3 followed by backti-
employing cathode current densities between 0.6 and tration of the remaining IO 3 with the
1:5kAm 2 a conversion to NaBH4 of 20 to 80% was I =I2ąS2O2 system.
3
claimed [4]. (ii) The semiquantitative silverąethylenediamine
More recently Shari�an together with Dutcher [5] (AgąEDA) method [10, 11]. This method is based
and Hale [6], respectively, extended the patent by on the reduction of Ag(I) by BH 4 in a 50%
Cooper to produce a variety of organic quaternary NaOH, 4% EDA solution. It was found to be a
ammonium and phosphonium borohydrides (i.e., very convenient spot test for BH 4 even for the
R1R2R3R4ąN�BH 4 and R1R2R3R4ąP�BH 4 , where nonaqueous samples analysed in the present
R1 4 can be alkyl, hydroxyalkyl or alkoxyl groups). work.
The above authors suggest as starting compounds a (iii) The crystal violet method [10] which was useful
number of boron oxides, such as metaborates, tetra- for nonaqueous samples.
borates B4O2 ą or perborates BO 3 ą. A current (iv) In addition to the above analytical techniques, a
9
e�ciency for sodium borohydride of 20% was new spot test was developed based on the re-
claimed after a 2 h electroreduction at 0:5kAm 2 on a duction of phosphotungstate PW12O3 ą by
40
nickel cathode when the catholyte was composed of BH 4 . It is well known that the Keggin type an-
10% by weight NaBO2 in 1 M NaOH [6]. In a similar ions (e.g., PW12O3 ą can be easily and reversibly
40
experiment a 25% current e�ciency for tetra- reduced to blueąviolet species, called heteropoly
methylammonium borohydride was achieved. blues [12]. This reaction was exploited to form
In spite of the industrial signi�cance of borohy- the basis of a convenient and simple test for BH 4
drides and the potential simplicity of the electro- detection.
0021-891X Ó 1998 Chapman & Hall 1147
1148 E. L. GYENGE AND C. W. OLOMAN
The following procedure was developed: to an of RaneyąNi on the stainless steel screen was per-
alkaline sample containing milligram amounts of formed according to the method described by Belot
BH 4 , about 0.2ą0.3 g of phosphotungstic acid et al. [13]. The electroplated RaneyąNi electrode was
H3PW12O40, Aldrich Inc.) was added. The Żask was activated before each run in 4 M NaOH.
swirled for about a minute followed by neutralization The nickel plate cathode was electropolished be-
of the sample with H2SO4 0.5 M. The neutral solution fore each run by anodically polarizing it in 60%
exhibited the characteristic blueąviolet colour of the H3PO4 for 2 min at 1 kA m 2 followed by sonication
heteropoly blue species formed by the BH 4 reduction in distilled water and methanol, respectively.
of PW12O3 . Additionally, in aqueous media two types of po-
40
The absorbance spectrum of the neutralized sam- rous cathodes were tested, i.e. RaneyąNi (see above)
ple (Fig. 1) was recorded in the range of 400 to and nickel boride (NiB, 35 mesh, 99% purity from
900 nm (scanning interval 1nm). A Novaspec spec- Cerac Inc.). The RaneyąNi was pretreated (activated)
trophotometer was employed (Pharmacia Biotech) before each run by digesting it in 4 M NaOH at 60 C
together with quartz Suprasil cuves (employable for 30 min [14].
wavelength range 200ą2500 nm) (Fisher Scienti�c The aqueous catholyte was a NaOH solution (0.1ą
Inc.). Distilled water was used as reference. A com- 3 M) containing various concentrations of dif-
puterized peak search (Novascan Software) per- ferent borate compounds such as NaBO2 (Aldrich
formed on the absorbance spectra given in Fig. 1, Inc.), H3BO3 (BDH Inc.) or borax Na2B4O7:10H2O
revealed the absorbance maximum occurring at Aldrich Inc:ą:
680 nm. The lowest BH 4 concentration which could The anode was a Pt mesh and the anolyte 1M
be detected by the above method was 10 4 M. NaOH.
The heteropoly blue species can be reoxidized by In organic media either a graphite rod A �
the oxygen present in the air to the colorless 3:3 cm2ą or an aluminum plate A � 5:1 cm2ą were
phosphotungstate form. Therefore, if the sample is employed as cathode. The catholyte consisted of tri-
not completely deoxygenated, the absorbance at methylborate BOCH3ą3, Aldrich Inc.) dissolved ei-
680 nm is decreasing with time, making the quanti- ther in ethylenediamine (Aldrich Inc.) or in a mixture
tative, spectrophotometric determination of BH 4 of hexamethylphosphoramide (Aldrich Inc.) and
impossible by the phosphotungstate method. ethanol. As supporting electrolyte in organic media,
either lithium chloride or lithium perchlorate were
used. The anolyte was 5 M LiOH and the anode a
2.2. Electrodes, electrolytes and apparatus
cylindrical Pt mesh.
employed for the electroreduction of borates
For the cathode plates the experimental apparatus
was an `H'-cell equipped with a cation exchange
The electroreduction of borates was studied on plate
membrane (Na�on� 324). The total catholyte volume
and �xed-bed cathodes in both aqueous and organic
was 150 ml.
media.
For the porous cathodes the �xed-bed cell presented
In aqueous media the following cathode plates
in Fig. 2 was used. As can be seen from Fig. 2 the basic
were used: nickel A � 8:2 cm2ą, amalgamated copper
framework of this cell is a glass funnel with a porous frit
A � 4:3 cm2ą, palladium A � 4:6 cm2ą, zinc A �
(total volume 60 ml, Pyrex�). The current feeder to the
5:7 cm2ą and RaneyąNi electroplated on a stainless
porous cathode was a circular Ni plate d � 4 cmą with
steel (316) screen (super�cial area 6:3 cm2ą. The
a Ni wire d � 0:05 cmą welded to it. The porous
RaneyąNi was purchased from Aldrich Inc. as a 50%
cathode material (i.e., RaneyąNi or NiB) was placed
slurry in water with a pore size of 50 l and a spe-
on the circular Ni current feeder up to a thickness be-
ci�c surface area of 80ą100 m2 g 1. The electroplating
tween 5 and 7 mm superficial area 12.6 cm2ą. The
cathode compartment was separated from the anode
compartment by a �ne porosity ceramic frit. As anode
a stainless steel (316) rod was employed.
The `funnel' electrochemical cell (Fig. 2) provided
a convenient solution for washing and rejuvenating
(activating) the porous cathodes without removing
the �ne particles from the cell.
The reference electrode for the aqueous media
experiments was a double junction saturated calomel
electrode.
All experiments were performed at room tem-
perature.
3. Results and discussion
The theoretical equation of metaborate BO 2 ą re-
Fig. 1. Absorbance spectra of the heteropoly blue species resulted
from the borohydride reduction of phosphotungstic acid. pH 7. duction to BH 4 in alkaline media is [15]:
TECHNICAL NOTE 1149
2 H2O � 2 e ! H2 � 2 OH 1ą
BO 2 � 4 H2 ! BH 4 � 2 H2O 2ą
The experiments performed are summarized in
Table 1. By employing several analytical methods and
performing `blank' experiments (Section 2.2) it was
found that none of the experiments presented in
Table 1, yielded any detectable amount of BH 4 .
Furthermore, it was observed that when the Raneyą
Ni bed was brought into contact with an alkaline
NaBH4 solution, strong hydrogen evolution oc-
curred. This indicates that the unpolarized RaneyąNi
catalyses the BH 4 decomposition [16], therefore it
cannot be employed as cathode for BH 4 electrosyn-
thesis. However, the same phenomena was not
observed on NiB.
The iodate method of BH 4 analysis (Section 2.2)
was found to be unreliable, giving erroneously high
borohydride concentrations. One of the reasons
Fig. 2. The �xed-bed, `funnel' batch electrochemical cell. Legend:
might be the insu�cient acidi�cation of the highly
(1) porous cathode, (2) anode, (3) separator (porous plug or cation
alkaline sample (e.g., 10% by weight NaBO2 in 1 M
exchange membrane), (4) reference electrode (SCE), (5) cathode
feeder plate, (6) cathode feeder rod, (7) porous plug, (8) glass NaOH). If the iodine titration with thiosulfate is
funnel, (9) glass tube.
performed at a pH insu�ciently acidic (e.g., the pH
is greater than 5 for a 10 3 N I2 solution [17]), IO is
generated as an intermediate and eight times less
BO 2 � 6 H2O � 8 e ! BH 4 � 8 OH
thiosulfate is consumed per one mole of iodine
E � 1:24 V vs SHE 1ą according to the following stoichiometry [17]:
S2O2 � 4 I2 � 10 OH ! 2 SO2 � 8 I � 5 H2O
In attempts to obtain the electrochemical reaction 3 4
given by Equation 1 two experimental strategies were
3ą
tested as follows: `indirect' electrocatalytic hydroge-
instead of the usual reaction
nation of metaborate and the `direct' electroreduction
of metaborate in alkaline media. Additionally, the 2 S2O2 � I2 ! S4O2 � 2 I 4ą
3 6
electroreduction of a boron compound in organic
Thus, being a backtitration, the less thiosulfate
media was investigated.
consumed can be wrongly interpreted as a certain
borohydride concentration. Furthermore, even in the
3.1. `Indirect' electrocatalytic hydrogenation case of su�cient acidi�cation, the iodate method of
BH 4 analysis failed for samples taken from the
This approach is based on the electroreduction of RaneyąNi and NiB �xed bed experiments (Table 1).
BO 2 in alkaline media on electrode materials which A black precipitate formed when KI was added to the
are hydrogenation catalysts, such as Ni, RaneyąNi, sample. The black precipitate rendered the thiosulfate
NiB, palladium and zinc. These experiments followed titration extremely inaccurate.
closely the experimental conditions indicated by the Because the electrocatalytic hydrogenation
patent literature [6]. Hale and Shari�an proposed the attempts of NaBO2 (Table 1) yielded no detectable
following mechanism for the electrochemical gener- amount of BH 4 , a number of experiments were per-
ation of BH 4 [6]: formed where the electrocatalytic hydrogen evolution
Table 1. Experimental conditions for the attempted electrocatalytic hydrogenation of borates
No. Cathode Catholyte Super�cial Cathode potential Reaction time
current density /V vs SCE /h
/kA m)2
1 Ni 10 wt % NaBO2, 1 M NaOH 0.50 )1.20 to )1.30 1
2 RaneyąNi electrodeposited ibid. 1.34 )1.30 3
on stainless-steel screen
3 RaneyąNi bed 10 wt % NaBO2, 1 M NaOH 1.60 )1.43 to )1.58 1
ibid. 3.50 )2.10
4 NiB bed 10 wt % NaBO2, 1 M NaOH 1.40 )1.70 to )2.12 3
5 Pd 10 wt % Na2B4O7.10H2O 0.10 )1.57 to )1.63 2
6 Zn 5 wt % NaBO2, 50 wt % K2CO3 3.5 ą 2
1150 E. L. GYENGE AND C. W. OLOMAN
Table 2. Experimental conditions for the attempted `direct' electroreduction of borates
No. Cathode Catholyte Super�cial Cathode potential Reaction
current density /V vs SCE time/h
/kA m)2
1 Amalgamated Cu 10 wt % NaBO2, 0.1 M NaOH 0.65 )2.18 to )2.26 1
2 Amalgamated Cu 10 wt % NaBO2, 2 M NaOH 7.50 )3.21 to )3.42 0.5
in TEAH*
5 wt % NaBO2, in TEAH 2.44 )2.87 to )3.12
3 Ni 20 wt % NaBO2, 1 M NaOH, 0.28 )1.35 to )1.41 1
0.1 M CTAB**
4 Ni 10 wt % NaBO2, 0.2 g dm)3 0.50 )1.60 to )1.70 2
thiourea
5 RaneyąNi electrodeposited 1.25 wt % H3BO3, 1 M NaOH, 1.60 )1.35 to )1.40 3
on stainless-steel screen 4 wt % (CH3)4NI
6 NiB bed ibid. 0.12 )1.30 4
7 Pd 10 wt % NaBO2, 3 M NaOH, 4.40 )1.90 to )2.01 1
0.2 g dm)3 thiourea
8 Zn 10 wt % NaBO2, 50 % K2CO3, 3.50 )2.67 to )2.87 1
0.2 g dm)3 thiourea
* -tetraethylammonium hydroxide 35 wt % solution in water.
** -cethyltrimethylammonium bromide.
was minimized in order to investigate the possibility 8 in Table 2), none of the experiments presented in
of a direct electrochemical reduction of borates to Table 2 gave any detectable amount of BH 4 .
BH 4 .
3.3. Electroreduction of a borate ester in organic media
3.2. `Direct' electroreduction of borates
in alkaline media Since the BH 4 electrosynthesis attempts in alkaline
aquous media were unsuccessful, the reduction of a
These experiments aimed at the suppression of the borate ester (i.e., trimethyl borate) in organic media
electrocatalytic hydrogen evolution, thereby `forcing' was investigated.
the possibility of a direct borate electroreduction. One of the most extreme reductions that one can
To increase the hydrogen evolution overpotential, perform is based on the so-called `solvated' electrons
besides selecting appropriate cathode materials such [23, 24]. In this procedure the commonly employed
as amalgamated copper, certain additives (i.e., qua- catholyte is either the hexamethylphosphoramide
ternaryammonium compounds and thiourea) were (HMPA)ąethanol mixture or certain amines (e.g.,
employed as well. Thiourea increases the hydrogen ethylenediamine, EDA [25]). Lithium salts (e.g.,
evolution overpotential by retarding the recombi- chloride or perchlorate) are the usual supporting
nation of the H atoms on the cathode surface [18ą electrolyte in these systems.
21]. As a consequence, strong H adsorption and Two experiments were performed under the above
surface hydride formation occurs on cathodes such conditions with graphite and aluminum cathodes
as Ni, Ni alloys and Pd [20, 22]. Quaternary (Table 3). There were no reducing species detected in
ammonium salts on the other hand, inhibit the either of the two experiments.
electrochemical step of the hydrogen evolution
mechanism [18, 21]. 4. Conclusions
The experimental conditions are summarized in
Table 2. Although signi�cant overpotentials vs. the There are a number of patents indicating the possi-
BO 2 =BH 4 standard potential were obtained in bility of electroreduction of borate compounds to
the presence of additives (e.g., entry no. 2, 7 and BH 4 with 20ą25% current e�ciency and 20 to 80%
Table 3. Experimental conditions for the attempted electroreduction of borates in organic media
No. Cathode Catholyte Super�cial Cell voltage Reaction
current density /V time/h
/kA m)2
1 Graphite 0.44 M B(OCH3)3, 1/2(mole/mole) HMPA*/ 0.08 50 3
ethanol, 0.1 M LiClO4
2 Al 1.32 M B(OCH3)3, 0.5 M LiCl, 0.09 30 2
0.1 M TBAHFP , in EDAą
* HMPA ą hexamethylphosphortriamide.
TBAHFP ą tetrabutylammonium hexaŻuorophosphate.
ą
EDA ą ethylenediamine.
TECHNICAL NOTE 1151
[4] B. H. Cooper, US Patent 3 734 842 (22 May 1973).
yield on electrocatalytic hydrogenation cathodes [4ą
[5] H. Shari�an and J. S. Dutcher, US Patent 4 904 357 (27
6]. In spite of the claims of the patent literature, our
Feb. 1990).
experiments aimed at the electroreduction of borates
[6] C. H. Hale and H. Shari�an, US Patent 4 931 154 (5 June
1990).
under both electrocatalytic hydrogenation and direct
[7] J. H. Morris, H. J. Gysling and D. Reed, Chem. Rev. 85
electroreduction conditions in alkaline media, did not
(1985) 51.
produce measurable amounts of BH 4 . Also, attempts
[8] M. V. Mirkin and A. J. Bard, Anal. Chem. 63 (1991) 532.
[9] D. A. Lyttle, E. H. Jensen and W. A. Struck, ibid. 24 (1952)
at the electroreduction of trimethyl borate under
1843.
`solvated electron' conditions generated no reducing
[10] Morton Performance Chemicals, Morton International,
species.
Danvers, MA, USA (1995).
[11] H. C. Brown and A. C. Boyd Jr., Anal. Chem. 27 (1955) 156.
The commonly employed iodate method of BH 4
[12] F. A. Cotton and G. Wilkinson, `Advanced Inorganic
analysis yielded false results in several cases. A new
Chemistry', 5th edition, John Wiley & Sons, New York
spot test for BH 4 detection was developed based on
(1988).
[13] G. Belot, S. Desjardins and J. Lessard, Tetrahedron Lett. 25
the reduction of phosphotungstic acid yielding the
(1984) 5347.
corresponding `heteropoly blue' species (absorbance
[14] T. Chiba, M. Okimoto, H. Nagai and Y. Takata, Bull.
maximum at 680 nm).
Chem. Soc. Jpn. 56 (1983) 719.
[15] S.-M. Park, Boron, in `Standard Potentials in Aqueous
Solution' (edited by A. J. Bard, R. Parsons and J. Jor-
Acknowledgement
dan), Marcel Dekker, New York (1985).
[16] E. Wiberg and E. Amberger, `Hydrides of the elements of
main groups IąIV', Elsevier, Amsterdam (1971).
The authors thank the Mechanical and Chemime-
[17] L. Kekedy, `Volumetric Analytical Chemistry (Titrimetry)',
chanical Wood-Pulps Network (one of the twelve
(in Hungarian), Dacia, Kolozsvar, Romania (1986).
Network of Centers of Excellence supported by the
[18] J. O'M. Bockris and S. U. M. Khan, `Surface Electro-
chemistry. A Molecular Level Approach', Plenum Press,
Canadian government) for the continuous interest
New York (1993)
and �nancial support for applied electrochemical re-
[19] A. C. D. Angelo and A. Lasia, J. Electrochem. Soc. 142
search. Also, sincere thanks to Dr Lawrence J. Gu-
(1995) 3313.
[20] Z. Szklarska-Smialowska and M. Smialowski, ibid. 110
ilbault, Vice-President R&D at Morton Performance
(1963) 444.
Chemicals, USA., for kindly supplying valuable in-
[21] T. Maoka and M. Enyo, Surf. Technol. 9 (1979) 147.
formation about borohydride and its analysis.
[22] H. Jarmolowicz and M. Smialowski, J. Catal. 1 (1962) 165.
[23] E. Kariv-Miller, R. I. Pacut and G. K. Lehman, Organic
Electroreductions at Very Negative Potentials, in `Topics
References
in Current Chemistry: Electrochemistry III' (edited by E.
Steckhan), Springer-Verlag, Berlin (1988).
[1] `Sodium Borohydride Digest', Morton International, Dan-
[24] J. Simonet, Electrogenerated Reagents, in `Organic Elec-
vers, MA, USA (1995).
trochemistry', 3rd edn (edited by H. Lund and M. Bai-
[2] Kirk-Othmer Encylcopedia of Chemical Technology, 4th
zer), Marcel Dekker, New York (1991).
edition, vol. 4 (edited by J. I. Kroschwitz), John Wiley &
[25] W. B. Schapp, R. E. Bayer, J. R. Siefker, J. Y. Kim, P. W.
Sons, New York (1992).
Brewster and F. C. Schmidt, Rec. Chem. Prog. 22 (1961)
[3] G. F. Hu� and R. M. Adams, US Patent 2 855 353, (7 Oct.
197.
1958).