Spontaneous hydrolysis of sodium borohydride in harsh conditions

i n t e rn a t i o n a l j o u r n a l o f h y d r o g e n en e r g y 3 6 ( 2 0 1 1 ) 2 2 4e2 3 3

Avai lab le at www.sc iencedi rect .com

journa l homepage : www.e lsev ie r . com/ loca te /he

Spontaneous hydrolysis of sodium borohydride in harshconditions

Jerome Andrieux a, Umit Bilge Demirci a,*, Julien Hannauer a, Christel Gervais b,Christelle Goutaudier a, Philippe Miele a

aUniversite Lyon 1, CNRS, UMR 5615, Laboratoire des Multimateriaux et Interfaces, 43 boulevard du 11 Novembre 1918,

F-69622 Villeurbanne, FrancebUniversite Paris 06, Laboratoire de Chimie de la Matiere Condensee Paris, CNRS, UMR 7574, College de France 11, place Marcelin Berthelot

F-75231 Paris Cedex 05, France

a r t i c l e i n f o

Article history:

Received 25 May 2010

Received in revised form

15 October 2010

Accepted 19 October 2010

Keywords:

Kinetics

Hydroxyborate anions

Sodium borohydride

Sodium metaborate

Spontaneous hydrolysis

* Corresponding author. Tel.: þ33372448403;E-mail address: Umit.Demirci@univ-lyon

0360-3199/$ e see front matter ª 2010 Profedoi:10.1016/j.ijhydene.2010.10.055

a b s t r a c t

The present study reports fundamental results about the spontaneous hydrolysis of

sodium borohydride NaBH4, which is a potential hydrogen storage material for small,

portable applications. The reaction (without stabilizing agent or catalyst) was carried out at

temperatures of 30e80 �C, initial NaBH4 concentrations of 0.63e6.19 mol L�1 (2.3e18.9 wt%),

and for unbuffered solutions, which are harsher experimental conditions than those

reported so far. The H2 evolution and the subsequent pH variation were observed to

determine the reaction kinetic parameters and characterize the hydrolysis intermediates,

i.e. the hydroxyborates BH4�z(OH)z�, by XRD, IR and 11B NMR. It was found that: (i) the

apparent activation energy of the reaction was 98 � 10 kJ mol�1 and the reaction order

versus the initial NaBH4 concentration was 0; (ii) all of the reactions BH4� / BH3(OH)�,

BH3(OH)� / BH2(OH)2�, BH2(OH)2

� / BH(OH)3� and BH(OH)3

� / B(OH)4� took place simulta-

neously; (iii) only 25 mol% of B(OH)4� and 75 mol% of BH4

� were found at 25% of conversion;

(iii) the hydroxyborates are very short-lived intermediates and only traces of BH3(OH)�

were detected.

ª 2010 Professor T. Nejat Veziroglu. Published by Elsevier Ltd. All rights reserved.

1. Introduction this reaction (1) [3e14]. Today, interest in the hydrolysis of

Sodium borohydride NaBH4 was first synthesized in the

1940se1950s by Schlesinger et al. [1] and soon attracted much

attention [2]. Stable in dry conditions, it spontaneously

(exothermically) hydrolyzes in the presence of water while

generating hydrogen:

BH�4 ðaqÞ þ 4H2OðlÞ/BðOHÞ�4 ðaqÞ þ 4H2ðgÞ (1)

In the 1950se1960s, most of the studies focused on

improving the knowledge of the kinetics and mechanisms of

fax: þ33372440618.1.fr (U.B. Demirci).ssor T. Nejat Veziroglu. P

NaBH4 is focused on applications, such that the objective is

typically to accelerate H2 release with the aid of metal cata-

lysts [15e22]. The spontaneous hydrolysis is less investigated

and it is even avoided by virtue of stabilizing agents such as

NaOH [23].

The spontaneous hydrolysis of NaBH4 is an acid-catalyzed

reaction (in-situ formed H3Oþ ions) [5,6]. Mochalov et al.

[10,11] proposed a detailed mechanism based on the substi-

tution of BeH bonds by BeOH bonds. The hydrolysis would be

a multistep process involving hydroxyborate intermediates

ublished by Elsevier Ltd. All rights reserved.

mailto:[email protected]

http://www.sciencedirect.com

http://www.elsevier.com/locate/he

http://dx.doi.org/10.1016/j.ijhydene.2010.10.055



i n t e r n a t i o n a l j o u rn a l o f h y d r o g e n en e r g y 3 6 ( 2 0 1 1 ) 2 2 4e2 3 3 225

BH4�z(OH)z� (with z from 1 to 3). For example, the reactions (2)

and (3) show the H3Oþ-catalyzed hydrolysis of BH4

� and the

formation of the first intermediate BH3(OH)� (i.e. z ¼ 1),

respectively:

BH�4 ðaqÞ þH3O

þðaqÞ/BH3ðaqÞ þH2OðlÞ þH2ðgÞ (2)

BH3ðaqÞ þOH�ðaqÞ/BH3OH�ðaqÞ (3)

The formation of BH3 from BH4� involves an activated

complex, which is HþBH4� or even BH5 (or H2BH3)

[5,6,8,10,11,14]. The reactions (2) and (3) are both rate deter-

mining steps.

The spontaneous hydrolysis of NaBH4 depends on experi-

mental conditions, e.g. temperature, solution pH and ratio

H2O/NaBH4 [5,6,24]. For example, a well-known empirical

relation (4) shows the dependence of the half-reaction time t1/2(in min) on the pH and the temperature T (in K) [10,11]:

log�t1=2;NaBH4

� ¼ pH� ð0:034 T� 1:92Þ (4)

To illustrate, based on this equation, hydrolysis should ach-

ieve a conversion of 50% in 134min at pH 10 at 15 �C, 28min at

pH 10 at 35 �C, 13271 min (i.e. >9 days) at pH 12 at 15 �C and

2773min (i.e. about 2 days) at pH 12 at 35 �C. This equationwas

established in the following experimental conditions:

7 < pH < 10 and 15 < T (�C) < 35. Even though it does not cover

pH ranges>10 and temperatures>35 �C, it is an indicative tool

that gives a satisfactory approximation of the effects of both

pH and temperature [15].

From this brief literature survey, it is clear that the kinetic

parameters reported so far were obtained at specific experi-

mental conditions. Most of the studies were carried out with

aqueous NaBH4 solutions at low concentrations (e.g.

0.01 mol L�1), which besides could be buffered [10,11,24].

Another point is that the kinetic data were calculated on the

few first percentages of NaBH4 converted [10,11,14]. Accord-

ingly, we studied the kinetics of the spontaneous hydrolysis

of NaBH4 at higher concentrations (up to about 6 mol L�1) and

high temperatures (30e80 �C). Furthermore, the solutions

were not buffered. Our objective was to contribute to a better

understanding of the reaction in different (more severe)

experimental conditions while attempting to identify the

reaction intermediates upon the evolution of 1 equiv H2.

Note that such severe conditions are much closer to the

conditions that should be used in technological applications,

even though catalysts will be used to enhance kinetics of H2

release.

Fig. 1 e Hydrogen evolutions obtained at various

temperatures (30, 50, 60, 70 and 80 �C) for an H2O/NaBH4

molar ratio of 9.

2. Experimental

Commercial NaBH4 (ACROS organics, 98% purity, powder,

n�CAS: 16940-66-2, average particle size of 200 mm) was used

as received. It was stored and handled in an argon-filled glove

box in order to avoid moisture contact and subsequent

hydrolysis. Distilled water was purged with argon to remove

oxygen.

The hydrolysis experiments were performed in a 20 mL

tube closed by a silicon stopper and placed in a thermostatic

bath. For the determination of the kinetic parameters the

overall system was considered as a non-steady state slurry

batch reactor. The H2O/NaBH4 molar ratio was equal to 9

(18.9 wt% of NaBH4 or about 6.18 mol L�1). To determine the

reaction order with respect to the initial NaBH4 concentration,

molar ratios of 13 (13.9 wt%), 18 (10.4 wt%), 27 (7.2 wt%), 36

(5.5 wt%) and 90 (2.3 wt%) were considered by varying the

NaBH4 mass. The NaBH4 content was below its solubility limit

(55 g per 100 g H2O) on the temperature rangewe used, namely

30e80 �C. Prior to any experiment, the NaBH4 solid was

transferred into the tube. In a typical experiment, water was

injected into the reactor with a needle placed directly inside

the bed of NaBH4. The injection needle was immediately

removed followingwater injection. An automated burette was

used to control the amount of water. The H2 produced was

collected and evacuated through a second needle, also inser-

ted in the tube through the stopper and connected to an

invertedwater-filled graduated cylinder. The H2 volumeswere

measured as a function of time by video monitoring the dis-

placed water as the reaction proceeded. The hydrolysis tests

were stopped when the H2 generation rate (HGR) was lower

than 0.0001 mL s�1, which is equivalent to a maximum

conversion value of 50e60%, depending on the experiment

conditions (the HGRs decreased because of the pH increase as

suggested by Equation (4)). A thermal probe placed inside the

NaBH4 bed measured the temperature during the reaction.

The NaBH4 conversion, 3 (%), was calculated as being the

ratio of the experimental volume of H2 released at a given time

to the expected maximum volume. The HGR (r, mL s�1) was

obtained by linear regression of the hydrogen evolution

curves. BecauseHGR changedwith the reaction conversion for

spontaneous hydrolysis, it was determined by taking the slope

of the curve at a given NaBH4 conversion range defined by

3 � 3%. For example, at 3 ¼ 20%, the range 17e23% was

considered and the data over this range were linearized. From

the as-obtained line (with R2 � 0.997), the HGR (i.e. the line

slope) was calculated.

The hydrolysis by-products were analyzed by X-ray

diffraction (Bruker D5005 powder diffractometer, CuKa radi-

ation (l ¼ 1.5406 A), Fourier transform infrared spectroscopy



Table 1 eHGRs (r, mL sL1) obtained for the H2O/NaBH4 molar ratio 9 (18.9 wt%), at various temperatures and determined forvarious conversions (3 ± 3%).

Temperature (�C) HGR (r, mL s�1) at various conversions

5% 10% 15% 20% 25% 30% 35% 40% 45% 50%

30 0.006 0.003 0.0025 0.0015 0.001 0.0005 ea ea ea ea

50 0.03 0.016 0.009 0.006 0.004 0.004 0.002 0.0012 0.0008 0.0008

60 0.08 0.03 0.021 0.014 0.008 0.006 0.005 0.003 0.003 0.0015

70 0.24 0.10 0.05 0.04 0.03 0.021 0.012 0.007 0.007 0.005

80 0.53 0.33 0.15 0.14 0.07 0.06 0.043 0.04 0.024 0.017

a The hydrolysis was stopped because of HGRs <0.0001 mL s�1.


(FT-IR, FTIR Nicolet 380) and 11B nuclear magnetic resonance

(either Bruker Avance500 MHz for solid compounds or Bruker

DRX 500 400 MHz for ions in aqueous solution). 11B MAS NMR

spectra of solid were recorded at 11.7 T on a Bruker Avance500

wide-bore spectrometer operating at 160.49 MHz, using

Fig. 2 e Plots of ln(r) as a function of TL1 for various conversion

activation energies as a function of conversion.

a Bruker 4 mm probe and a spinning frequency of the rotor of

14 kHz. Spectra were acquired using a spin-echo q-t-2q pulse

sequence with q ¼ 90� to overcome problems of probe signal

and recycle delay of 1s was used. Chemical shifts were refer-

enced to BF3(OEt)2 (d ¼ 0 ppm).

s, i.e. (a) 5%, (b) 10%, (c) 30%, and (d) 50%, and (e) apparent



Fig. 3 e Hydrogen evolution at conversions of 20% (a) and

40% (b) for various H2O/NaBH4 molar ratios, at 80 �C, and (c)

determination of the reaction order versus the NaBH4

concentration at 80 �C and for conversions of 20 ± 3% and

40 ± 3%.


3. Results and discussion

3.1. Kinetic study

The NaBH4 spontaneous hydrolysis was studied from 30 to

80 �C. The molar ratio H2O/NaBH4 was kept at 9. Fig. 1 shows

the H2 evolutions obtained at 30, 50, 60, 70 and 80 �C. As

expected, the higher the temperature, themore the hydrolysis

extent (in terms of HGRs and conversions at a given time). This

is also shown by the data given in Table 1. Moreover, the

curves had similar profiles; for a given temperature, the HGRs

decreased with time, i.e. with the reaction conversion. As

suggested by Equation (4), this was due to an increase of the

solution pH because of the formation of the basic hydrox-

yborates BH4�z(OH)z� [10,11]. In fact, it was observed that at

conversions of around 50% the HGRs were very low, at this

point the experiments were stopped.

The data in Table 1 were then exploited to plot

InðrÞ ¼ fð1=TÞ and the apparent activation energy was calcu-

lated. This was done for HGRs obtained at various conver-

sions, i.e. 5 � 3%, 10 � 3%, 15 � 3%, 20 � 3%, 25 � 3%, 30 � 3%,

35 � 3%, 40 � 3%, 45 � 3%, and 50 � 3%. For example, Fig. 2

(aed) depicts some of the as-obtained linear regressions. It is

important to note that for a given conversion the solution

characteristics (pH, products concentration and so on) should

have been similar, with the only different parameter being the

temperature. Fig. 2(e) shows the apparent activation energies

as a function of the conversion. The apparent activation

energy is 98 � 10 kJ mol�1 over the range 30e80 �C (lit.

[4e9,12,13]: 100 � 35 kJ mol�1 at low conversions, low

temperatures, i.e. < 35 �C, and for buffered solutions).

The reaction order versus the initial NaBH4 concentration

was determined for various H2O/NaBH4 molar ratios (i.e. 9, 13,

18, 27, 36 and 90; 6.18, 4.27, 3.09, 2.08, 1.54 and 0.63 mol L�1;

18.9, 13.9, 10.4, 7.2, 5.5 and 2.3 wt%) at 80 �C. The H2 evolution

curves were analyzed at both 20 � 3% and 40 � 3% (Fig. 3(a,b));

at these conversions, the H2 evolution was quite linear, and

the HGRs were deduced by linear regression. Table 2 reports

the as-determined HGRs. Three observations stood out: (i) the

HGRs were quite similar, i.e. about 0.17mL s�1, at a conversion

of 20% and irrespective of the molar ratios; (ii) a similar

observation was reported at a conversion of 40% (HGRs of

0.04e0.07 mL s�1); (iii) at a given reaction time, the lower the

H2O/NaBH4 molar ratio (i.e. the higher the NaBH4 concentra-

tion), the higher the volume of H2 released. The HGR datawere

then used to plot their logarithmas a function of the logarithm

of the initial NaBH4 concentration (Fig. 3(c)). A reaction order

versus the initial NaBH4 concentration of zero was obtained at

20 � 3% and 40 � 3%. In other words, the spontaneous

hydrolysis is independent on the initial NaBH4 concentration

for concentrations up to 6.18 mol L�1. To the best of our

knowledge, reaction orders versus the initial NaBH4 concen-

tration in these experimental conditions have not been

reported to date, which does not make any comparison

possible. It is noteworthy that a reaction order of 0 was

obtained for lower conversions and temperatures and for

buffered solutions [4e9,12,13].

To summarize the results discussed heretofore, it

mainly stands out that the apparent activation energy of the

spontaneous hydrolysis determined over the range 30e80 �C is

98 � 10 kJ mol�1 and is consistent with the energies obtained

at lower temperatures. Given that the apparent activation

energy determined experimentally includes the energies of

every occurring reaction, one may assume that all of the

reactions BH4� / BH3(OH)�, BH3(OH)� / BH2(OH)2

�,BH2(OH)2

� / BH(OH)3� and BH(OH)3

� / B(OH)4� take place

simultaneously. Another assumption might be that the reac-

tion BH4� / B(OH)4

� preferentially takes place and predomi-

nates. To highlight this, the hydrolysis slurry was analyzed

upon the evolution of 1 equiv H2.

3.2. Effect of pH on hydrolysis

According to Mochalov et al. [10,11], the solution pH has

a significant effect on the spontaneous hydrolysis kinetics.We

therefore followed the pH evolution in the course of the

spontaneoushydrolysisofNaBH4at 30 �C (Fig. 4(a)). ThepHfirst

increased quite rapidly from about 9 to 10.8 (in about 16 min).



Table 2 e HGRs (r) for various molar ratios and for conversions of 20 ± 3% and 40 ± 3%.

H2O/NaBH4 molar ratio [NaBH4]0 (mol L�1) NaBH4]0 (wt%) Conversion of 20 � 3% Conversion of 40 � 3%

r (mL s�1) Dra r (mL s�1) Dra

90 0.63 2.3 0.20 0.06 0.07 0.03

36 1.54 5.5 0.16 0.05 0.07 0.03

27 2.08 7.2 0.16 0.03 0.08 0.02

18 3.09 10.4 0.16 0.02 0.07 0.01

13 4.27 13.9 0.17 0.04 0.06 0.01

9 6.18 18.9 0.14 0.02 0.04 0.05

a With Dr, the experimental error on r.


Then, the pH still increased but less rapidly. Actually the pH

increase per unit of time was less and less high as the hydro-

lysis proceeded. It increased up to 11.7 after 30000 s of hydro-

lysis (i.e. >8 h). The profile of the pH evolution was similar to

that of the H2 evolution. This clearly suggests that the solution

pH strongly influences the hydrolysis kinetics. Nevertheless,

the apparent activation energy is constant within the conver-

sion range 0e50% and the hydrolysis is independent on the

initial NaBH4 concentration. Hence, one may assume that the

pH has no effect on the reaction mechanisms.

The NaBH4 spontaneous hydrolysis is a complex reaction

involvingat least5different reactionspecies, fromBH4� toB(OH)4

�,which play an important role in the evolution of the pH solution.

Mochalov et al. suggested a stepwise hydrolysis of NaBH4 and

showed that the half-reaction times of NaBH3(OH), NaBH2(OH)2and NaBH(OH)3 were lower than that of NaBH4 [10,11]:

Fig. 4 e Evolution of the solution pH (experimental and

calculated) as a function of time for an H2O/NaBH4 molar

ratio of 9 at 30 �C (a) and difference between the

experimental and calculated pH values (b).

log t1=2;NaBH3ðOHÞ ¼ pH� ð0:027 T� 0:357Þ (5)
� �
log�t1=2;NaBH2ðOHÞ2

� ¼ pH� ð0:027 T� 0:384Þ (6)

log�t1=2;NaBHðOHÞ3

� ¼ pH� ð0:024 T� 4:00Þ (7)

The equations (4)e(7) were established under the following

experimental conditions: 7 < pH < 10, 15 < T (�C) < 35, and

buffered solutions. The rate constants of the reactions

BH4� / BH3(OH)�, BH3(OH)� / BH2(OH)2

�, BH2(OH)2� / BH

(OH)3� and BH(OH)3

� / B(OH)4� were found as 5.3 � 107,

3.6 � 1011, 10.1 � 107, 5.7 � 1010 mol�1 min�1, respectively. A

Fig. 5 e XRD patterns of the solids recovered at 5, 10, 15, 20,

25% of conversion from top to bottom (a) and XRD pattern

of the product obtained by the reaction of 1 equiv. H2O with

4 NaBH4 (b); the stars show the peaks attributed to NaBH4

(ICDD 00-009-0386) while the others are assigned to NaB

(OH)4 (ICDD 01-081-1512); the broad peaks within the range

2q [ 13�e28� is due to the amorphous film used to prevent

the sample from moisture.




constant rate of 5.2 � 107 mol�1 min�1 was also determined

for the reaction BH4� / B(OH)4

�. In fact, BH4� can hydrolyze

into B(OH)4� according to two routes: a stepwise one via the

formation of BH3(OH)� or a direct one, the rate constants of

the respective reactions being equivalent. It has been also

reported that the concentrations of the BH3(OH)� and BH

(OH)3� ions were small (0.03e0.05 mol%) in the course of

the hydrolysis whereas the concentration of BH2(OH)2�

conversion w26 mol%. Molina Concha et al. [25] showed the

formation of BH3(OH)� through heterogeneous hydrolysis of

NaBH4 and the presence of both BH3 and BH2 as stable

intermediates by in-situ IR measurements.

According to Ruman et al., it should be possible to

synthesize BH3(OH)� through the hydrolysis under ambient

conditions of one BH4� by one H2O in an excess of THF [26]. Its

formation was evidenced by 1H NMR and the spectrum was

similar to that of NaBH4 as it showed similar BeH coupling

region but it was different in the region containing the BH

resonances. An interesting point in Ruman et al’s conclusion

Fig. 6 e IR spectra of NaBH4 and solids recovered at 5, 10,

15, 20, 25% of conversion from top to bottom (a) and IR

spectrum of the product obtained by the reaction of

1 equiv. H2O with 4 NaBH4 (b).

is that they assumed all of the BH4� ions underwent a one-step

hydrolysis and thus all led to the formation of BH3(OH)�.UnlikeMochalov et al’s studies [10,11], the NaBH4 solutionwas

not buffered. If the discrepancies in the experimental condi-

tions are not considered, Ruman et al’s work is in a way in

contradiction with Mochalov et al’s.

In order to highlight the discrepancy reported above, we

tentatively calculated a theoretical pH evolution on the basis

of the pKa of the couple B(OH)3/B(OH)4�, which is 9.2 [27]. This

couple is involved in the stepwise hydrolysis of NaBH4 [10,11]:

BðOHÞ3ðaqÞ þH2OðlÞ/BðOHÞ�4 ðaqÞ þHþðaqÞ (8)

Even though considering B(OH)3 as the only, first acid will

overestimate the pH, it was assumed that the pH evolution

was only dependent on the formation of B(OH)4� from B(OH)3

and the following relation was applied, with B(OH)3 being

a weak acid:

pH ¼ 7þ 0:5 pKaþ 0:5loghBðOHÞ�4

i(9)

Fig. 4(a,b) shows the calculated evolution. Three observa-

tions stood out: (i) the profile of the calculated evolution was

quite similar to that of the experimental one; (ii) consistent

with the acid/base couple considered and its pKa, the calcu-

lated pH was overestimated (DpH of about 1) at the beginning

of the hydrolysis (1e3% of conversion) but interestingly DpH

decreased to 0.1 at 17e18%; (iii) for conversions higher than

22% (i.e. t> 20000 s), the calculated pHwas almost equal to the

experimental pH. Observations (ii) and (iii) suggest therefore

that the acid/base couple(s) involving the first hydrolysis

intermediates such as e.g. BH3(OH)� would have a pKa value

lower than that of the couple B(OH)3/B(OH)4�, and that the

solution pH would be controlled by B(OH)4� at conversion

Fig. 7 e Solid 11B RMN spectrum of the product obtained by

the reaction of 1 equiv. H2O with 4 NaBH4 (a) and proton

coupled 11B RMN spectrum of this product in aqueous

(90 wt% H2O D 10 wt% D2O) solution.




�22%. In other words, it is important to identify the reaction

intermediates that are involved in the beginning of the

hydrolysis.

3.3. Hydrolysis intermediates identification

To make clear the pH evolution, we focused our work on

characterizing the hydrolysis intermediates at conversions up

to 25%. The following tests were performed. The NaBH4

spontaneous hydrolysis was stopped at various conversions,

i.e. 5, 10, 15, 20 and 25%, by dipping the tube in liquid air

(<�140 �C). The slurry froze within 10 s, which is negligible in

comparison to the timescale required for completion of

spontaneous NaBH4 hydrolysis. The reaction was thus effec-

tively quenched. The frozen slurry was dried under vacuum at

Fig. 8 e Proton decoupled (top) and coupled (bottom) 11B NMR spe

of 25%.

�48 �C for 48 h (to remove all water and prevent any hydro-

lysis when the sample was warmed) and then handled in

an argon-filled glove box. The solids were analyzed by XRD

(Fig. 5(a)). NaBH4 (ICDD 00-009-0386) was clearly identified at

each conversion. It was the sole phase identified whereas

a likely peak at around 2q ¼ 32�, characteristic of borates, had

been especially regarded. No borate was found, which could

be due an amorphous state [28]. Since the NaBH4 hydrolysis

products have been reported as being stable up to 170 �C [29],

our solids were heated at 150 �C for 4 h under Ar but such

process was unsuccessful in crystallizing the borates.

As a further test, the solids were then analyzed by IR

(Fig. 6(a)). Three observations stood out. The spectrum recor-

ded forNaBH4was confirmed [30,31]. TheNaBH4 print, i.e. nmax/

cm�1 1114, 2220, 2293, 2341, 2357 and 2438 (BeH), was clearly

ctra of the hydrolysis by-product recovered at a conversion



Fig. 9 e Spontaneous hydrolysis of NaBH4 and

intermediates effect on pH over the conversion range

0e25%.


visible in each of the other spectra, showing that most of the

reactant had not reacted. New bands, at nmax/cm�1 �500 d

(BeO), 700e900 n(BeO), 1200 d(BeOeH), 1330 n(BeO), 1550e1650

d(HeOeH) and 3300e3500 n(OeH) [25,32], were observed for the

partly hydrolyzed NaBH4, their respective contribution

appearing to increase with the NaBH4 conversion. From such

observations, it is in fact quite difficult to assert the formation

of one hydroxyborate compound over another one. Tenta-

tively our IR spectra might be compared to spectra previously

reported and some/most of the bandsmight be assigned to e.g.

NaBH3(OH) [29], NaBO2 [33,34], NaB(OH)4 [35], or NaBO2.2H2O

[29,33,36]. In Ref. [25], bands at 970 and 1180 cm�1 were

attributed to BH2 and BH3, respectively. Nevertheless, none

clearly matched and one may propose that the IR spectra

indicate the formation of a tetra-coordinated boron but the

presence of both BeH (especially from unreacted NaBH4) and

BeOeH stretching modes impedes a clear identification of

BH4�z(OH)z� or B(OH)4

�.To succeed in identifying the compounds NaBH4�z(OH)z

that could be found in the hydrolysis medium at a conversion

of 25%, we attempted to synthesize NaBH3(OH) according to

the procedure recently reported by Ruman et al. [26]. Typically

1 equiv H2O per 1 NaBH4 weremixed in extra dry THF under Ar

and the slurry was aged for 40 days while being stirred. Then,

it was dried under vacuum for 24 h, transferred in the glove

box, ground and analyzed by XRD. The XRD pattern is given in

Fig. 5(b). The peaks could be assigned to both NaBH4 and NaB

(OH)4. It is noteworthy that, some synthesis attempts showed

unreacted NaBH4 and the presence of peaks assigned to NaB

(OH)4. Nevertheless, as NaBH3(OH) is not listed in our XRD

patterns database, we did not discard its formation. The

synthesized product was then analyzed by IR (Fig. 6(b)). The

spectrum supported the formation of NaB(OH)4, matching IR

bands reported elsewhere [29]. As our observations contradict

those of Ruman et al. [26], we completed the analysis by per-

forming 11B solid state MAS NMR. The spectrum is given in

Fig. 7(a). Two signals were observed at 2.2 and �41.8 ppm

(difference of 44.0 ppm), corresponding to NaB(OH)4 and

NaBH4, respectively [37]. Furthermore, we prepared an

aqueous (90 wt% H2O and 10 wt% D2O) solution of this

compound to be analyzed by 11B NMR.Water was injected into

the sample-filled vial only few seconds before the analysis.

Despite the H2 evolution, a spectrum not decoupled from

proton in Fig. 7(b) was obtained. Once more, two signals were

observed at 5.1 and �38.5 ppm (difference of 43.6 ppm), cor-

responding to B(OH)4� and BH4

�, respectively. The spectrum

consisted of a singlet (NaB(OH)4) and a quintet (NaBH4). The

JBH spinespin coupling constant for BH4� was 80.6 Hz (lit.

[12,13,26]: 82.0, 81.0 Hz). By focusing on d z �10 ppm, it was

possible to distinguish within the background noise a negli-

gibly small signal, likely a quartet that could be assigned to

BH3(OH)�, which could have formed by hydrolysis of the dis-

solved solid. As a result, XRD, IR and NMR strongly suggested

the formation of NaB(OH)4 rather than that of NaBH3(OH). In

other words, when 1 equiv H2O is intended to hydrolyze 1

NaBH4, it is in fact 25% of the NaBH4 that hydrolyzes whereas

75% does not hydrolyze. This implies that at a conversion of

25% the specie controlling the solution pH is NaB(OH)4.11B RMN was also used to analyze the slurry recovered at a

conversion of 25 � 1%. In this case, despite a low H2 evolution,

the 11B NMR collecting was done for the solution to keep the

slurry state unchanged. The temperature was set at 5 �C to

minimize the H2 evolution. Fig. 8(a) shows the proton decou-

pled spectrum. It showed 2 signals at �38.4 and 4.8 ppm

(difference of 43.2 ppm), characteristic of BH4� and B(OH)4

� [37].

A very small signal was observed at �9.9 ppm, ascribed to

BH3(OH)�. Assuming a conversion of NaBH4 of 25%, the inte-

gration gave a relative content in BH3(OH)� of 0.4 mol%, rather

in agreement with Mochalov et al.’s observations [10,11]. No

other hydroxyborate was detected. Guella et al. [37] investi-

gated thehydrolysis intermediates for Pd-catalyzedhydrolysis

of NaBH4. In a typical 11B proton decoupled NMR spectrum, no

compound other than BH4� and B(OH)4

� was present, what was

attributed to the low symmetry of the hydroxyborates struc-

ture leading to a broadening of 11B signals. Another reason

could be the presence of the Pd catalyst, which accelerated the

hydrolysis and so decreased the lifetime of very short-lived

intermediates such as BH3(OH)� [10,11]. It is noteworthy that,

for uncatalyzed hydrolysis of NaBH4, the main thermody-

namically stable borate is NaB(OH)4 in experimental condi-

tions different from ours [28,38,39].

Fig. 8(b) shows the 11B spectrum not decoupled from

proton, which is rather consistent with that reported by

Gardiner and Collat [12,13]. The spectrum consisted of

a singlet (NaB(OH)4), a quartet (NaBH3(OH)) and a quintet

(NaBH4). The JBH spinespin coupling constants were calcu-

lated. The JBH value for NaBH3(OH) was 87.6 Hz (lit.

[12,13,26,37]: 87.5, 82.0, 81.0 Hz). The JBH value for NaBH4 was

80.6 Hz (lit. [12,13,26]: 82.0, 81.0 Hz). The presence of the

hydroxyborates BH2(OH)2� and BH(OH)3

� was especially scruti-

nized by exploring the shifts within the range defined by the

signals of the borates BH3(OH)� and B(OH)4�, a shielding being

expected. As shown in Fig. 8, no BH2(OH)2� or BH(OH)3

� were

detected. In other words, when NaBH3(OH) was synthesized

by reacting 1:1H2O with NaBH4 or by quenching the NaBH4

hydrolysis at a conversion of 25%, mainly NaB(OH)4 was

formed with traces of NaBH3(OH).

4. Conclusion

The spontaneous hydrolysis of NaBH4 was studied under

experimental conditions thataremuchmoresevere than those




reported to date (i.e. T¼ 30e80 �C, [NaBH4]0¼ 0.63e6.19mol L�1,

solutions not buffered). Under these conditions, an apparent

activationenergyof98�10kJmol�1andareactionorderversus

the initial NaBH4 concentration of 0 were found. Besides,

a parallelity in the evolution of the solution pHwith that of the

NaBH4 conversion was observed.

The hydrolysis intermediates at conversions up to 25%

were recovered by dipping the hydrolysis slurry in liquid air

(<�140 �C) and then dried under vacuum. The as-obtained

solids were analyzed by XRD, IR and 11B NMR. The three

characterization techniques gave the same results; they evi-

denced the formation of 25 mol% of B(OH)4� and 75 mol% of

BH4� at a conversion of 25%. This was consistent with the

kinetic data. It was besides confirmed that BH3(OH)�, which is

the first hydrolysis intermediate, is a very short-lived inter-

mediate. Traces of it were detected. No other hydroxyborate

was detected by 11B NMR.

As a result, one concludes that (i) the direct reaction

BH4� / BH3(OH)� preferentially occurred under our experi-

mental conditions and (ii) the B(OH)4� anions controlled the

solution pH at conversions �22%. Fig. 9 depicts schematically

our observations. Notably, our study has confirmed that at

basic pH, the H2 release through spontaneous hydrolysis is

harshly hindered.

Acknowledgements

The present studywasmainly funded by the ‘Cluster Energies’

of ‘Region Rhone-Alpes, France’. It was also funded by the ANR

project BoraHCx. The authors are grateful to Mrs. Caroline

TOPPAN (Univ. Lyon 1, CCRMN) andMr. OlivierMajoulet (Univ.

Lyon 1, LMI) for some of the 11B NMR spectra and the subse-

quent valuable discussions.

r e f e r e n c e s

[1] Schlesinger HI, Brown HC, Finholt AE, Gilbreath JR,Hoekstra HR, Hyde EK. Sodium borohydride, its hydrolysisand its use as a reducing agent and in the generation ofhydrogen. J Am Chem Soc 1953;75:215e9.

[2] Demirci UB, Miele P. Sodium tetrahydroborate as hydrogen/energy carrier, its history. C R Chim 2009;12:943e50.

[3] Kilpatrick M, McKinney CD. Kinetics of the reaction oflithium borohydride in aqueous solution. J Am Chem Soc1950;72:5474e6.

[4] Pecsok RL. Polarographic studies of the oxidation andhydrolysis of sodium borohydride. J Am Chem Soc 1953;75:2862e4.

[5] Davis RE, Swain CG. General acid catalysis of the hydrolysisof sodium borohydride. J Am Chem Soc 1960;82:5949e50.

[6] Davis RE, Bromels E, Kibby CL. Boron hydrides. III. Hydrolysisof sodium borohydride in aqueous solution. J Am Chem Soc1962;84:885e92.

[7] Stockmayer WH, Miller RR, Zeto RJ. Kinetics of borohydridehydrolysis. J Phys Chem A 1961;61:1076e7.

[8] Jolly WL, Mesmer RE. Exchange of hydrogen between waterand hydroborate ion. J Am Chem Soc 1961;83:4470e1.

[9] Mesmer RE, Jolly WL. The hydrolysis of aqueous hydroborate.Inorg Chem 1962;1:608e12.

[10] Mochalov KN, Khain VS, Gil’manshin GG. Kinetic studies onintermediate steps of BH4

� hydrolysis. Kinet Katal 1965;6:541e4.

[11] Mochalov KN, Khain VS, Gil’manshin GG. Generalizedscheme for the hydrolysis of the borohydride ion anddiborane. Dokl Akad Nauk 1965;162:402e6.

[12] Gardiner JA, Collat JW. The hydrolysis of sodiumtetrahydroborates. Identification of an intermediate. J AmChem Soc 1964;86:3165e6.

[13] Gardiner JA, Collat JW. Kinetics of the stepwise hydrolysis oftetrahydroborate ion. J Am Chem Soc 1965;87:1692e700.

[14] Kreevoy MM, Hutchins JEC. H2BH3 as an intermediate intetrahydroborate hydrolysis. J Am Chem Soc 1972;94:6371e6.

[15] Amendola SC, Sharp-Goldman SL, Janjua MS, Spencer NC,Kelly MT, Petillo PJ, et al. A safe, portable, hydrogen gasgenerator using aqueous borohydride solution and Rucatalyst. Int J Hydrogen Energy 2000;25:969e75.

[16] Metin O, Ozkar S. Hydrogen generation from the hydrolysisof sodium borohydride by using water dispersible,hydrogenphosphate-stabilized nickel(0) nanoclusters ascatalyst. Int J Hydrogen Energy 2007;32:1707e15.

[17] Hung AJ, Tsai SF, Hsu YY, Ku JR, Chen YH, Yu CC. Kinetics ofsodium borohydride hydrolysis reaction for hydrogengeneration. Int J Hydrogen Energy 2008;33:6205e15.

[18] Liu BH, Li Q. A highly active CoeB catalyst for hydrogengeneration from sodium borohydride hydrolysis. Int JHydrogen Energy 2008;33:7385e91.

[19] Oronzio R, Monteleone G, Pozio A, De Francesco M, Galli S.New reactor design for catalytic sodium borohydridehydrolysis. Int J Hydrogen Energy 2009;34:4555e60.

[20] Akdim O, Demirci UB, Miele P. More reactive cobalt chloridein the hydrolysis of sodium borohydride. Int J HydrogenEnergy 2009;34:9444e9.

[21] Prosini PP, Gislon P. Water consumption during solid statesodium borohydride hydrolysis. Int J Hydrogen Energy; 2010.doi:10.1016/j.ijhydene.2010.07.094.

[22] Ding XL, Yuan X, Jia C, Ma ZF. Hydrogen generation fromcatalytic hydrolysis of sodium borohydride solution usingCobalteCoppereBoride (CoeCueB) catalysts. Int J HydrogenEnergy; 2010. doi:10.1016/j.ijhydene.2010.07.030.

[23] Hua D, Hanxi Y, Xinping A, Chuansin C. Hydrogenproduction from catalytic hydrolysis of sodium borohydridesolution using nickel boride catalyst. Int J Hydrogen Energy2003;28:1095e100.

[24] Jensen EH. A study on sodium borohydride with specialreference to its analytical application in organic chemistry.Copenhagen: Nyt. Nordisk Forlag; 1954.

[25] Molina Concha B, Chatenet M, Maillard F, Ticianelli EA,Lima FHB, de Lima RB. In situ infrared (FTIR) study of themechanism of the borohydride oxidation reaction. PhysChem Phys Chem 2010;12:11507e16.

[26] Ruman T, Anna Ku�snierz A, Jurkiewicz A, Le�s A, Rode W. Thesynthesis, reactivity and 1H NMR investigation of thehydroxyborohydride anion. Inorg Chem Commun 2007;10:1074e8.

[27] CRC handbook of chemistry and physics. 81st ed. Boca Raton:CRC Press; 2000.

[28] Stepanov N, Uvarov V, Popov N, Sasson Y. Study of by-product of NaBH4 hydrolysis and its behavior at a roomtemperature. Int J Hydrogen Energy 2008;33:7378e84.

[29] Goubeau VJ, Kallfass H. Die reaktion natriumborhydrid andwasser. Z Anorg Allg Chem; 1959:160e9.

[30] Harvey KT, McQuaker NR. Infrared and Raman spectra ofpotassium and sodium borohydride. Can J Chem 1971;49:3272e81.

[31] Somer M, Acar S, Koz C, Kokal I, Hohn P, Cardoso-Gil R, et al.a- and b-Na2[NH4][NH2]: two modifications of a complex




hydride in the system NaNH2-NaBH4; syntheses, crystalstructures, thermal analyses, mass and vibrational spectra. JAlloys Compd 2009;491:98e105.

[32] Jun L, Shuping X, Shiyang G. TF-IR and Raman spectroscopicstudy of hydrated borates. Spectrochim Acta 1995;51A:519e32.

[33] Miller FA, Wilkins CH. Infrared spectra and characteristicsfrequencies of inorganic ions. Anal Chem 1952;24:1253e61.

[34] Molina Concha B, Chatenet M, Coutanceau C, Hahn F. In situinfrared (FTIR) study of the borohydride oxidation reaction.Electrochem Commun 2009;11:223e6.

[35] Janda R, Heller G. IR- and Ramanspektren isotop markiertertetra- und pentaborate? Spectrochim Acta 1982;36:997e1001.

[36] Kanturk A, Sari M, Piskin S. Synthesis, crystal structure anddehydration kinetics of NaB(OH)4$2H2O. Korean J Chem Eng2008;25:1331e7.

[37] Guella G, Zanchetta C, Patton B, Miotello A. New insights onthe mechanism of palladium-catalyzed hydrolysis of sodiumborohydride from 11B NMR measurements. J Phys Chem B2006;110:17024e33.

[38] Marrero-Alfonso EY, Davis TA, Matthews MA. Minimizingwater utilization in hydrolysis of sodium borohydride: therole of sodium metaborate hydrates. Int J Hydrogen Energy2007;32:4717e22.

[39] Marrero-Alfonso EY, Gray JR, Davis TA, Matthews MA.Hydrolysis of sodium borohydride with steam. Int J HydrogenEnergy 2007;32:4723e8.



Spontaneous hydrolysis of sodium borohydride in harsh conditions

Documents