- Email: [email protected]
A study of the anomalous behaviour of the glass electrode in solutions containing hydrofluoric acid
JOTJR%i_4L
128
OF
ELECTROAN_4LYTICAL
CHEJIISTRY
STUDY OF THE ANOMALOUS BEHAVIOUR OF THE GLASS ELECTRODE IN SOLUTIONS COXTAINING HYDROFL’UORIC ACID
A
E:.
S0RENSEN
The
Damish
AND
Aionric
(Received
October
I-_
Emcrgy
LUFlDG_&%RD Comwzisszox
Research
Establiskwzent.
RisG
/Dennzauk/
28th. 1964)
Although the research of recent years has brought about a considerable extension of the applicability of the glass electrode, its behaviour in solutions containing free HF has attracted little attentionlp’ considering the special action of this acid on &SS.
In this laboratory a glass electrode has been used for some time as a convenient device for controlling the addition of HF to aqueous solutions_ This mode of operation was the incentive to the present work, in which we examine the variation of the glasselectrode potential when the outer glass surface is exposed to dilute solutions of HF. EXPERIMENTS
A Radiometer glass-calomel electrode assembly, type GK combination with a Radiometer pH-meter, model 22, and a Ovarian
The perspex electrode (Fig_ I). fraction
The
vessel is made so as to allow a continuous volume of liquid contained in the vessel
small
of a second
through
the two
introduction
holes,
ZOZB, graphic
is used
in
recorder_
flow of the test solutions can be replaced within a
connected
with
independent
feed pumps. In order to eliminate variations in H+ activity, all the solutions used are 0.1 hy which has been soaked in in HCL The temperature is kept at 21~. A glass electrode 0.1 N HCl overnight is placed in the cell and treated alternately with pure 0.1 N HCL and-with O-I mole/l of Hl? in 0-r LV HCl, both applied at a rate of 0.5 ml /sec. During this process the potential difference is registered by the recorder as shown in Fig. 2. The-potent&al drop caused by the first contact with HF is seen to be different from subse+ent ones. -This initial behaviour is the more marked the longer the electrode has b&n s&king m_ XlF-free - aqueous ‘solution. During the_ periods of regeneration the potential;cmve shows a steep rise, which is followed by a moderately ascending part converging towar&_ the norn-ial electrode potential (Fig. 2; II and IV)_ As the data giv@ be&w presupp&z.s the sensitivity of a recently HF-treated glass electrode, the r&e;ner&oti_time G kept within 3 min. and reproducible results are obtained by inlro+rci;lg the HF solution at the Gurie stage every time. E-xperim&ts with var$i&g [HFJ gave a series of potential 3s. time cu~~=es. A %le&on df khese-isshown m Fig. 3. They are seein to be composed essentially of two : _ :- - _ _. J- Eliidroanal.
_, .Ch_-,
-
-
,I
'g:_(Ig65) mxd3-~33
BEHAVIOUR
OF
THE
GLXSS
ELECTRODE
LS
I.
HF
CONTAINING
129
I
-3 fror$lpe: mp Fig.
SOLUTIONS
time
Glass-calomel
electrode
assembly
and
electrode
Fig. a. Graph of potential vs. time during pre-treatment overnight in O-I LN HCL Periods I. III and V, flow of and IV, regeneration with pure 0.1 _~\rHCl_
m
minutes
vessel. of a glass electrode which has 0.1 mole/l of HF in 0.1 .N HCI;
been kept periods II
-----_ I D
E
F
-----em--we--
1 umt =I5 set
a02
Log [HF]
‘t”N’>
Ql
Fig_ =J_ Potential change of a pre-treated and regenerated glass electrode on contact with a series of HF in 0.1 iV HCI soIutions. The starting points correspond to the beginning of periods III and V in Fig. 2. (-4). 0-1 N; (ES). 0.04 A’: (C), 0.02 IV; (D), 0.0125 N; (E), o 0078 N~r56p_p_m_ HF: (F). o.oo3g Nm 78 p-p-m_ HF_-The dashed line indicates the approximate limit of the initial drop. Fig-
+
Log
(0~ +
#l) us. log
[HFI
([HFI
>
a.01
N). I- ELecbroand-
them..
9 (1965)
128-133
r3o
E.
50RESSE?G,
T.
L‘LTSDGAARD
parts, namely a rapid drop, which is nearly 5.5 mV for the curves representing [HFj above o-or N, and a further decrease The latter dominates at higher [HF], where it is almost linear. At lower [HF] the slope declines while the linearity becomes less pronounced. If the slope expressed in mV/sec is denoted by cy, a constant, 8, can be determined in such a way that the plot of log (LY+ #?) W. log [HFJ is strictly linear in the lower part of the concentration range (Fig_ 4). The meaning of t5’will be discussed later. It appears that for each [I-IF] above o-or N a new constant potential is established. If the difference between the normal glass-electrode potential in 0.1 X HCI and the HF-depressed potential is called dE ESF and plotted against log [HF], a curve is obtained which deviates little from a straight line (Fig_ j)_ This result formally corresponds to an HIS-electrode function : Em-Ea=wherez
=
~4%
g
log [HF']
,
BEHAVIOUR
OF
THE
Fig. 6. Log (dEEIF)
GLASS
log
VS.
ELECTRODE
[HF~
([HF;
IN
<
o.oI
SOLUTIOXS
CONTAINIXG
HF
’ 131
x)_
DISCUSSIOK
According to BOKSAY et al.3 it is assumed that practically only protons participate in the electrode processes taking place in the hydrated surface layer. Voltage anomalies may then arise from the exchange of some of the proton acceptors for others to which the protons are bound with altered forces. When a hydrated glass surface is exposed to dilute hydrofluoric acid, a re&tion takes place which can be expressed in the simplified form [Si,,i]
(OlQ
t
y HF
+
[S&i]
(OH),-,
FY +
y l&O,
where _Xandy represent average numbers. If we assume that the voltage deviation of the electrode is proportional to the degree of surface fluorination, dE/dt will be closely related to the reaction rate. The terms, u and fl, of Fig. 4 are interpreted as the rates of the initial net reaction and the reverse reaction, respectively. This is confirmed by the fact that the magnitude of #? is
very similar to the average slope of the regeneration curve after the steep rise. Thus (cx+ p) represents the forward reaction rate. It is seen from Fig. 4 that (z + 8) is proportional part equals glass-electrode
to [HF]? in the lower concentration range as the slope of the Linear curve z-o_ This can be explained by supposing that the observed change of the potential
follows
upon
the uptake
of 2 F per Si-atom,
where the second
step is-the rate-determining step. Finally the reverse react& rate reaches a magnitude equal to that of the. forward reaction rate (compare the steep rise of the regeneration curve), and a stationary state is established corresponding to the constant value of the glass-electrode potential_ The negative
deviation
is in agreement
with
the assumption
that
the substitu-
tion bf F.- fer._OH- increases the dissociation constant of silicic acid. Tfre relation presented in Fig. 6. shows a .fonnal similarity with
a Freundlich adsorption isotherm/It may well be .ssumed that the- potential drop in -question is This is supported by the low [HF]. --due to’ &3sorptio&of .JfIFby: the ‘&ES .&face. : s.ufficient..to .pk&ce the effect; an&the instantane,ous charkter of the.latter. :.:
-.
_-
:.
-. :-- .: .,.._, ...I ,;.,__:-: _., ..: .: .: ., :y..:
.:.. . ., -. .~
..._. .... .. ._
._
::
:
J.. Ekctroa~s~~.Gh&z.;.g
(196.5
rzSrrj.3
In a routine chemical operation a TOO/, soiution of HF was added to a uranium ieach &.g-w containing chiefly Na&XI-r, /&(SO& and Si0~. aq. When the formation of fluoride conmlexes ceased, the potential of an immersed glass electrode underwent a characteristic -hange, indicating a slight excess of free HF. Howewr, the amount added did not zfj;roe with calculations based upon an analysis of the liquor and certain stoichiwuetric assmptions, It was therefore decided to perfcwm. sirnifar experiments with separate compaunds of elements known to form fluoride complexes. To 50 ml of each of the fo~~o~~~g solutions, contained in a polystyrene beaker equipped with a nzagzmlic stirrer, z .M NF was added at a rate of 5 ml/ruin :
IT-------
I
TI
bides- of_ Hi
i
pee- -
i
rmsie _of t&rant
tvroles at HF per m&r c+fIfib-ant
(c) 0.2 &f ZrOCl-) in 0.x LV HCI; (d) 0.2 M AlCls in 0.1 N HCl ; (c) 0.x M X1+0& in 0.1 AT HCl; (f) silicic-acid solution (0.75% SiOz) in 0.1 N HCl; fg) (f) -j- z KC1 per 23% potential as a Figures 7(aP b, c) show the variations of the glass-electrode function of the amount of I-IF added. The initial rise is due to Iiberation of II+ during the complex formation_ The drop later is indicative of free HF in the mixture. X semi-quantrtative determmation can be made on comparison with Fig. 3_ It should be noted, however, that when the [HFJ exceeds 0.05 molt/l, no further details can be obtained by this method. In Fig_ 7a Fe 3+ is introduced in the form of perchlorate in order to avoid cumplexing competition from anions other than F-_ In the presence of ZrO’- the HF concn. is negligible until the stoichiometric amount has been added. This property of ZrCPf makes it the preferred titrant fur F-. In Fig. 7b the difference between (d) and (e) is probably due to the formation of a mixed comple_x of Al 3fP SOS’- and F-. In fact this agrees with experience from the manufacture of synthetic cryohtc, where sulphate in the raw materials inevitably gets into the product. * -62-j ensured by the solubility product In Fig. 7c the effect of K- is a low [!%I of IGSiFs. Even then a considerable surplus of HF is needed to convert SiOs. aq. entirely into SiFs”-. XCKSOWLEDGEXEXT
The authors have benefited greatly from discussions, T. R~sEK~~~ERG.
particularly
with C. 32.
LISDERSTR@JSI-LANG~~~~~~I~~~
-4 defined surface hydration of the glass electrode is secured by pre-treatment with 0.x _M HF followed by rinsing with pure 0-1 Ar HCE for a few minutes. On subsequent contact with 0.1 X HCl containing HF, the electrode potential shows a change which is determined by the [HI;)_ The immediate reaction is an adsorption of HF by the glass surface- This is followed. at fHF]higher than 0.01 N, by the substitution of two F- for OH- per Si atom. ?%rith increasing HF the attack on the Si-O-5 bonds becomes severe, but it can be toierated to a considerable degree because the newly formed surface is identical with the previous one_ Some applications of the glass electrode in titrations with HF are suggested in an appendix.
128
OF
ELECTROAN_4LYTICAL
CHEJIISTRY
STUDY OF THE ANOMALOUS BEHAVIOUR OF THE GLASS ELECTRODE IN SOLUTIONS COXTAINING HYDROFL’UORIC ACID
A
E:.
S0RENSEN
The
Damish
AND
Aionric
(Received
October
I-_
Emcrgy
LUFlDG_&%RD Comwzisszox
Research
Establiskwzent.
RisG
/Dennzauk/
28th. 1964)
Although the research of recent years has brought about a considerable extension of the applicability of the glass electrode, its behaviour in solutions containing free HF has attracted little attentionlp’ considering the special action of this acid on &SS.
In this laboratory a glass electrode has been used for some time as a convenient device for controlling the addition of HF to aqueous solutions_ This mode of operation was the incentive to the present work, in which we examine the variation of the glasselectrode potential when the outer glass surface is exposed to dilute solutions of HF. EXPERIMENTS
A Radiometer glass-calomel electrode assembly, type GK combination with a Radiometer pH-meter, model 22, and a Ovarian
The perspex electrode (Fig_ I). fraction
The
vessel is made so as to allow a continuous volume of liquid contained in the vessel
small
of a second
through
the two
introduction
holes,
ZOZB, graphic
is used
in
recorder_
flow of the test solutions can be replaced within a
connected
with
independent
feed pumps. In order to eliminate variations in H+ activity, all the solutions used are 0.1 hy which has been soaked in in HCL The temperature is kept at 21~. A glass electrode 0.1 N HCl overnight is placed in the cell and treated alternately with pure 0.1 N HCL and-with O-I mole/l of Hl? in 0-r LV HCl, both applied at a rate of 0.5 ml /sec. During this process the potential difference is registered by the recorder as shown in Fig. 2. The-potent&al drop caused by the first contact with HF is seen to be different from subse+ent ones. -This initial behaviour is the more marked the longer the electrode has b&n s&king m_ XlF-free - aqueous ‘solution. During the_ periods of regeneration the potential;cmve shows a steep rise, which is followed by a moderately ascending part converging towar&_ the norn-ial electrode potential (Fig. 2; II and IV)_ As the data giv@ be&w presupp&z.s the sensitivity of a recently HF-treated glass electrode, the r&e;ner&oti_time G kept within 3 min. and reproducible results are obtained by inlro+rci;lg the HF solution at the Gurie stage every time. E-xperim&ts with var$i&g [HFJ gave a series of potential 3s. time cu~~=es. A %le&on df khese-isshown m Fig. 3. They are seein to be composed essentially of two : _ :- - _ _. J- Eliidroanal.
_, .Ch_-,
-
-
,I
'g:_(Ig65) mxd3-~33
BEHAVIOUR
OF
THE
GLXSS
ELECTRODE
LS
I.
HF
CONTAINING
129
I
-3 fror$lpe: mp Fig.
SOLUTIONS
time
Glass-calomel
electrode
assembly
and
electrode
Fig. a. Graph of potential vs. time during pre-treatment overnight in O-I LN HCL Periods I. III and V, flow of and IV, regeneration with pure 0.1 _~\rHCl_
m
minutes
vessel. of a glass electrode which has 0.1 mole/l of HF in 0.1 .N HCI;
been kept periods II
-----_ I D
E
F
-----em--we--
1 umt =I5 set
a02
Log [HF]
‘t”N’>
Ql
Fig_ =J_ Potential change of a pre-treated and regenerated glass electrode on contact with a series of HF in 0.1 iV HCI soIutions. The starting points correspond to the beginning of periods III and V in Fig. 2. (-4). 0-1 N; (ES). 0.04 A’: (C), 0.02 IV; (D), 0.0125 N; (E), o 0078 N~r56p_p_m_ HF: (F). o.oo3g Nm 78 p-p-m_ HF_-The dashed line indicates the approximate limit of the initial drop. Fig-
+
Log
(0~ +
#l) us. log
[HFI
([HFI
>
a.01
N). I- ELecbroand-
them..
9 (1965)
128-133
r3o
E.
50RESSE?G,
T.
L‘LTSDGAARD
parts, namely a rapid drop, which is nearly 5.5 mV for the curves representing [HFj above o-or N, and a further decrease The latter dominates at higher [HF], where it is almost linear. At lower [HF] the slope declines while the linearity becomes less pronounced. If the slope expressed in mV/sec is denoted by cy, a constant, 8, can be determined in such a way that the plot of log (LY+ #?) W. log [HFJ is strictly linear in the lower part of the concentration range (Fig_ 4). The meaning of t5’will be discussed later. It appears that for each [I-IF] above o-or N a new constant potential is established. If the difference between the normal glass-electrode potential in 0.1 X HCI and the HF-depressed potential is called dE ESF and plotted against log [HF], a curve is obtained which deviates little from a straight line (Fig_ j)_ This result formally corresponds to an HIS-electrode function : Em-Ea=wherez
=
~4%
g
log [HF']
,
BEHAVIOUR
OF
THE
Fig. 6. Log (dEEIF)
GLASS
log
VS.
ELECTRODE
[HF~
([HF;
IN
<
o.oI
SOLUTIOXS
CONTAINIXG
HF
’ 131
x)_
DISCUSSIOK
According to BOKSAY et al.3 it is assumed that practically only protons participate in the electrode processes taking place in the hydrated surface layer. Voltage anomalies may then arise from the exchange of some of the proton acceptors for others to which the protons are bound with altered forces. When a hydrated glass surface is exposed to dilute hydrofluoric acid, a re&tion takes place which can be expressed in the simplified form [Si,,i]
(OlQ
t
y HF
+
[S&i]
(OH),-,
FY +
y l&O,
where _Xandy represent average numbers. If we assume that the voltage deviation of the electrode is proportional to the degree of surface fluorination, dE/dt will be closely related to the reaction rate. The terms, u and fl, of Fig. 4 are interpreted as the rates of the initial net reaction and the reverse reaction, respectively. This is confirmed by the fact that the magnitude of #? is
very similar to the average slope of the regeneration curve after the steep rise. Thus (cx+ p) represents the forward reaction rate. It is seen from Fig. 4 that (z + 8) is proportional part equals glass-electrode
to [HF]? in the lower concentration range as the slope of the Linear curve z-o_ This can be explained by supposing that the observed change of the potential
follows
upon
the uptake
of 2 F per Si-atom,
where the second
step is-the rate-determining step. Finally the reverse react& rate reaches a magnitude equal to that of the. forward reaction rate (compare the steep rise of the regeneration curve), and a stationary state is established corresponding to the constant value of the glass-electrode potential_ The negative
deviation
is in agreement
with
the assumption
that
the substitu-
tion bf F.- fer._OH- increases the dissociation constant of silicic acid. Tfre relation presented in Fig. 6. shows a .fonnal similarity with
a Freundlich adsorption isotherm/It may well be .ssumed that the- potential drop in -question is This is supported by the low [HF]. --due to’ &3sorptio&of .JfIFby: the ‘&ES .&face. : s.ufficient..to .pk&ce the effect; an&the instantane,ous charkter of the.latter. :.:
-.
_-
:.
-. :-- .: .,.._, ...I ,;.,__:-: _., ..: .: .: ., :y..:
.:.. . ., -. .~
..._. .... .. ._
._
::
:
J.. Ekctroa~s~~.Gh&z.;.g
(196.5
rzSrrj.3
In a routine chemical operation a TOO/, soiution of HF was added to a uranium ieach &.g-w containing chiefly Na&XI-r, /&(SO& and Si0~. aq. When the formation of fluoride conmlexes ceased, the potential of an immersed glass electrode underwent a characteristic -hange, indicating a slight excess of free HF. Howewr, the amount added did not zfj;roe with calculations based upon an analysis of the liquor and certain stoichiwuetric assmptions, It was therefore decided to perfcwm. sirnifar experiments with separate compaunds of elements known to form fluoride complexes. To 50 ml of each of the fo~~o~~~g solutions, contained in a polystyrene beaker equipped with a nzagzmlic stirrer, z .M NF was added at a rate of 5 ml/ruin :
IT-------
I
TI
bides- of_ Hi
i
pee- -
i
rmsie _of t&rant
tvroles at HF per m&r c+fIfib-ant
(c) 0.2 &f ZrOCl-) in 0.x LV HCI; (d) 0.2 M AlCls in 0.1 N HCl ; (c) 0.x M X1+0& in 0.1 AT HCl; (f) silicic-acid solution (0.75% SiOz) in 0.1 N HCl; fg) (f) -j- z KC1 per 23% potential as a Figures 7(aP b, c) show the variations of the glass-electrode function of the amount of I-IF added. The initial rise is due to Iiberation of II+ during the complex formation_ The drop later is indicative of free HF in the mixture. X semi-quantrtative determmation can be made on comparison with Fig. 3_ It should be noted, however, that when the [HFJ exceeds 0.05 molt/l, no further details can be obtained by this method. In Fig_ 7a Fe 3+ is introduced in the form of perchlorate in order to avoid cumplexing competition from anions other than F-_ In the presence of ZrO’- the HF concn. is negligible until the stoichiometric amount has been added. This property of ZrCPf makes it the preferred titrant fur F-. In Fig. 7b the difference between (d) and (e) is probably due to the formation of a mixed comple_x of Al 3fP SOS’- and F-. In fact this agrees with experience from the manufacture of synthetic cryohtc, where sulphate in the raw materials inevitably gets into the product. * -62-j ensured by the solubility product In Fig. 7c the effect of K- is a low [!%I of IGSiFs. Even then a considerable surplus of HF is needed to convert SiOs. aq. entirely into SiFs”-. XCKSOWLEDGEXEXT
The authors have benefited greatly from discussions, T. R~sEK~~~ERG.
particularly
with C. 32.
LISDERSTR@JSI-LANG~~~~~~I~~~
-4 defined surface hydration of the glass electrode is secured by pre-treatment with 0.x _M HF followed by rinsing with pure 0-1 Ar HCE for a few minutes. On subsequent contact with 0.1 X HCl containing HF, the electrode potential shows a change which is determined by the [HI;)_ The immediate reaction is an adsorption of HF by the glass surface- This is followed. at fHF]higher than 0.01 N, by the substitution of two F- for OH- per Si atom. ?%rith increasing HF the attack on the Si-O-5 bonds becomes severe, but it can be toierated to a considerable degree because the newly formed surface is identical with the previous one_ Some applications of the glass electrode in titrations with HF are suggested in an appendix.