JIIIIHILIll

ELSEVIER

Journal of Power Sources 53 (1995) 9%107

Gas evolution and performance assessment of submarine lead/acid batteries

J.B. L a k e m a n 1
DRA West Drayton, Kingston Lane, West Drayton, Middx. UB7 9QB, UK

Received 22 August 1994; accepted 9 September 1994

Abstract

The gas-evolution rates from new and end-of-life lead/acid submarine cells at both open-circuit and under float-charge have been examined, together with the discharge performance of the cells. The evolved gas is found to comprise hydrogen and oxygen only. The open-circuit and float-charge hydrogen-evolution rates vary logarithmically with temperature. The enthalpy of activation for open-circuit hydrogen evolution is ~ 70 kJ mo1-1 for both old and new cells. The open-circuit hydrogenevolution rate per Ah of rated capacity for new cells is between 0.00636 and 0.105 cm3 h -1 All -1 at STP at 20-50 C, respectively. For old cells, a > three-fold increase is observed, the evolution rate ranges from 0.0237 to 0.361 cm3 h-1 Ah-1 at STP at 20-50 C, respectively. On float-charge, using an uncompensated float voltage, the hydrogen-evolution rate of new cells is between 0.0166 and 0.241 cm3 h-~ Ah-~ at STP at 20-50 C, respectively. For old cells, the rates increase to 0.0298--0.903 cm3 h-I A h - I at STP at 20-50 C, respectively. In contrast, the electrical discharge performance of the cells decreases with age. This suggests that the active area for hydrogen evolution increases with age, but that the active area for electrical discharge decreases. No correlation has been established between hydrogen evolution in individual cells and electrical discharge performance for cells of the same age. The self-discharge rates of the negative plates at 20 C are 14 and 50% per year for new and old cells, respectively.
Keywords: Open-circuit; Float-charge; Gas evolution; Self-discharge; Lead/acid batteries; Submarines

1. Introduction On-charge gas evolution, and the associated charging efficiency of lead/acid cells, can be used as an indicator of cell condition. T h e gas evolved derives largely from the electrolysis of water on overcharge, with the components hydrogen and oxygen produced in a fairly constant ratio [1]. Low evolution of gas is associated with better charge retention and lower maintenance (topping up) [2]. On open-circuit, hydrogen evolution is associated with self-discharge of the negative plate, and oxygen with the positive. Therefore, the composition of the evolved gas could vary. On float-charge, it can be assumed that the evolved gas will derive both from self-discharge mechanisms and from the electrolysis of water, In well-ventilated enclosures, the gassing of aqueous electrolyte batteries does not present a serious hazard. By contrast, in confined spaces, such as in a submarine, ~Tel.: 0895 456411.

gas evolution can pose a serious threat to vehicle safety, and must be kept to a minimum when the vehicle is submerged. It is, therefore, of critical importance that the rate of hydrogen evolution, particularly, is known at all stages of the submarine battery duty cycle, in order that adequate gas-management systems can be designed and implemented. The increased usage of valve-regulated lead/acid batteries ( V R L A ) has also highlighted the need to minimize gas evolution, and a great deal of research has already been conducted (see, for example, Refs. [3-6]). T h e data obtained on very large submarine cells can be extrapolated back to give more accurate information about much smaller systems, for which precise data for gas evolution on open-circuit and float-charge, are not available because of the small amounts of gas evolved. A report of on-charge gas-evolution experiments on submarine cells has been published previously [1]. The work discussed here concerns the more operationally critical scenario, when the submarine is submerged,

British Crown Copyright 1994/DRA -- published with the permission of the Controller of Her Britannic Majesty's Stationery Office SSDI 0378-7753(94)02004-M

CHARGE BREAKER SHOWN OFF POSITIVZ B U S B ~

DISCItAItGE BREAKER SHOWN ON ~

I(:oMPu~E. ~O~OL'QU'"ENTt
CHARGE
I I

1
F=7 CURkENT
Ism,'m [ % " WATER..,,~ war1

. . . . . . . . . . . . . . . . . . . . . . . . . . . ~
I~" w a r

L ~v LOAI
WAn

I I

-I

r~-AIc ~

SHUNT

YT

L O A D > LO~ % LOAD.

VOLTAGE I
'

ER

2"

1"

CELLS I
UNDER I - CU3IIIENT comet_ I /

I I

KEYBOARD
LOAD I LOAD I LOAD I LOAD BItEAKE~ BREAKERI BRJEAKERI BREAKEI

E)[~

WATER I WATER I WATER I w.qBt

I / HIGHF~WER TJLLOAD CoNrrRo[

'
I
I I J LOAD SANK CURRENT

L'~ "

[I COMPUTERAND FACILITY [~SPECIFIC1/0 BOARDS

I
I

SHUNIS

I t I
I

=~ *~

.......................
[

--2 OFF
,

,SO~N 16 OFF

SYSTEM AND TH~ COMPUTER 15 oFF

I

2 OFF
A A

]1 ~ I

,~ "~
~ I VOLT^GE^~CURRENT~ ''

CHARGE

BATTERY

DISCHARGE

. . . .

[CONFIRMATION OFTHE

CURRENTI~3100 1123.sVOLTAGEI11760CURRENTI

SIGNALS{ lCHARGER CONTROL ]

BREAKERCOMMANU

IWATERLOAD THERMOSTAT [ IBREAKERSTATE'AND THE I lFACILITYCONTROL AND I

[ LOW AND HIGIIPOWER I 16 BIT DRIVEFOR TIlE

/ 1

L~

I O[FI?UT~~;'~; DIGITAL

DIGITAL INPUTS

I ANALOG[IEINPUTS ~

~ ~:~

ALL MONITORINGAND CONTROL

SIGNALSARE STEEREDTI~ROUGH I]O

-BREAKERSIf THE AU'II3 LOOp STOPS

1II T
WATCHDOG TIMER TUI~NSOFFTHE CHARGE AND DISCHARGE

Fig, 1. Schematic diagram of the micro-computer controlled submarine battery test facility at D R A West Drayton.

J.B. Lakeman /Journal of Power Sources 53 (1995) 99-107

101

and the battery is either on open-circuit or on floatcharge. The effect of service life on hydrogen-evolution rate is examined by testing both new cells and cells at the end of service. The temperature dependency of gas evolution, and the relationship between evolution rate and electrical discharge performance are also examined.

sured continuously using a paramagnetic oxygen analyzer; the hydrogen content was derived by difference. 2.2. Electrical performance The discharge performance of the two batteries was measured using a recently developed micro-computer controlled, 16-bit, 0.5 MW battery test facility. A schematic of the facility is shown in Fig. 1. Discharge/ charge profiles of up to 45 steps of constant power, current and resistance on discharge, and constant current or voltage on charge, can be programmed in any combination. Discharge is achieved by using resistive loads; the maximum current of 16 000 A is resolved to ~ 0.25 A. Control precision is limited by the seven current-measuring shunts to better than 0.2% of reading. Charging is carried out using a thyristor-controlled analogue supply. The discharge current is controlled by arranging the resistive elements into two, 8-bit binary pattern groups: one high power, the other low power. The high-power load bank is shown schematically in Fig. 2. Power MOSFET heat-sink modules are used to switch the individual resistive elements of both banks. For very high power tests, the load is supplemented by four water-cooled resistive loads. A typical five-stage constant power discharge profile on the 10-cell battery is given in Fig. 3. 3. Results and discussion Fig. 4 illustrates the variation, at open-circuit, of the hydrogen-evolution rate (in cm 3 h -1 cell -1) at STP
MAXIMUM BUSBAR VOLTAGE 30V

2. Experimental Two batteries of submarine cells were tested. In each case the negative grids were made from a binary lead-calcium alloy, and the positive grids were Pb-3wt.%Sb-l.5wt.%Sn-0.05wt.%Se [7]. One battery of ten cells was new and was received in the drycharged state. The cells were filled with sulfuric acid and cycled to the rated capacity; the rated performance and charging procedure has already been described [1,2]. The second battery, of eight cells, was selected at random from an end-of-life submarine battery, 2.1. Gas evolution Cells were incubated in a thermostatically-controlled water tank at temperatures ranging from ~ 11 to 53 C. Gas-evolution rates were measured either at opencircuit following a full gassing charge, or on a floatcharge of 2.20 to 2.22 V/cell. The rate of gas evolution was measured continuously with electronic mass flowmeters. The gas composition was analyzed initially using a quadrupole mass spectrometer, in order to establish that the only significant components were oxygen and hydrogen. Gas compositions were subsequently mea-

[~

POSITIVE BUSBAR ONE OUTOF THE 255 RESISTORS

11.5 VOLT SUPPLY

0.4 OHM >1.5 KW RESISTOR CONTROL COMPUTER CONTROL SWITCH IIEATS1NK MODULE
TIIREE MOSFET TRANSISTORS CONNECTED ON H E A T S I N K
..... I-I

MANUAL

!A U T O

BIT DRIVE LED

CONTROL SWITCH COMMON

l ~INDICATION

CURRENT
SHUNT

ONE OF THE EIGHT BIT D R I V E SIGNALS FOR TIlE IIIGH POWER LOAD BANK (DIS I o D | )

ONEOF TIlE SEVEN c ~ SHUb'I~ CONNECTEDTOTHE HIGHPOW1EItLOADIIANK

[
Fig. 2. Schematic diagram of the high-power load bank.

NEGATIVE BUSBAR

102
20,

J.B. Lakeman / Journal of Power Sources 53 (1995) 99-107

-30 power

1,.~ -~
m
) 1 8 . ol

I
~
vo

L~

~0
-35

-45

-so ,-ss o ~

17'

16,

, -60 ' -65

15. , -70 14 0 I 2000 I 4000 I 6000

Time / seconds

I 8000

I 10000

-7 12000

Fig. 3. A five-step constant-power discharge test on 10-cell battery of new ceils.

1200

1000

U D. 3: D. 0 0 U) o

800

600

400-

200

i 40 i 80
Time / hours

!
120

i
100

i 200

Fig. 4. Average hydrogen-evolution rate vs. time on open-circuit after charge for end-of-life submarine cells at 38 C.

1600

~ 300

250 0 1200 0 0 0 a. "I 2 0 0 a. (.1 0

.100
,00,
oxygen 50

0
0

I 40

I 80
Time / hours

I
120

I
160

0 200

Fig. 5. Average total gas and oxygen-evolution rates vs. time on open-circuit after charge for end-of-life submarine cells at 38 C.

(SCCPHPC), against time elapsed since a full gassing charge for end-of-life submarine cells at 38 C. Initially, there is a decrease in flow rate as the residual gas from the previous charge is released. The evolution rate then progressively increases over ~ 30 h to reach

the maximum. The corresponding traces for total gas, and for oxygen, are displayed in Fig. 5. These traces achieve a maximum in ~11 h, following which the oxygen-evolution rate gradually declines. The reason for the decline in oxygen evolution is unclear, but must

J.B. Lakeman / Journal of Power Sources 53 (1995) 99-107 400.

103

300
t~ a. -r a.
m

2OO

u.

100

0 0

i 40

i 80 Time I hours

i 120

i 160

i 200

Fig. 6. Average hydrogen-evolution rate vs. time on open-circuit after charge for new submarine cells at 38 C.

500 U ~. 400 "I(1. 0

U

100 ~ 90 80 ~ n 70 O. "1" 60 50 oxygen ~ 40 m= t:: 30 20 0 I 40 I 80 Time I hours I 120 I 160 10 0 200

0

m ~

300

: c~ ~o 200
I--

100 0 0

Fig. 7. Average total gas and oxygen-evolution rates vs. time on open-circuit after charge for new submarine cells at 38 C. Table 1 Temperature, hydrogen content, and average hydrogen-evolution rate for end-of-life cells on open-circuit

more slowly in the case of new cells. The average, steady-state, hydrogen flow-rate for old and new cells at 38 C was found to be 1040 and 315 SCCPHPC, respectively. This indicates that the open-circuit hydrogen-evolution rate has increased by more than threefold over the life of the cells. Results at other temperatures are listed in Tables 1 and 2 for old and new cells, respectively. In each case, the hydrogen content of the evolved gas is always
Table 2 Temperature, hydrogen content, and average hydrogen evolution rate for new cells on open-circuit Temperature H2
(%) 98.9 98.9 98.9

Temperature (c)
22 29

H2 (%)
99

Hz flow rate

(SCCPHPC)
184 491 1307 1026
3707 3848

11.4

98.5 98.3

89

29
38 38 53 53.3

98.2
94.5 96.7 89.8 90.5

536

r e f l e c t a r e d u c t i o n in t h e a c t i v e a r e a f o r s e l f - d i s c h a r g e at t h e p o s i t i v e , r a t h e r t h a n a n i n c r e a s e in t h e r a t e o f o x y g e n r e d u c t i o n at t h e n e g a t i v e .

(c)

H2 flow rate (SCCPHPC)

74 94 1 16 210

Fig. 6 illustrates the variation in the average hydrogenevolution rate at open-circuit with time for new submarine cells at 38 C. Fig. 7 shows the corresponding traces for total gas and for oxygen. Comparison of Figs. 4 and 5 with Figs. 6 and 7 reveals that the hydrogenevolution rate rises, and the oxygen-evolution rate falls,

21 24.3 28.7 34.9

98.9

51.5

38.3 39.7 44.3

96.8

95.3 98.8 97.2

361 312
1286
557

104

J.B. Lakeman / Journal of Power Sources 53 (1995) 99-107

90% or above. This confirms previous observations that the rate of self-discharge of the positive is significantly lower than that of the negative [8]. In the case of old cells, there is some indication that the hydrogen content decreases with increasing temperature. The open-circuit, hydrogen-evolution rate is found to vary logarithmically with temperature, and the results of Tables i and 2 are plotted in Figs. 8 and 9, respectively, with the regression equations indicated. F r o m the equations, the open-circuit, hydrogen-evolution rate per Ah of rated capacity for new cells was found to be between 0.00636 and 0.105 cm 3 h -1 Ah -1 at STP, at 20-50 C, respectively. For old cells, the corresponding evolution rates are between 0.0237 and 0.361 cm 3 h -1 Ah -1 at STP.

Table 4
Hydrogen-evolution Temperature

data for new cells on float-charge

H2 M e a n cell Float

H2 flow

(c)

(SCCPHPC)

(%)

voltage (v) 2.213 2.219 2.209 2.219 2.218 2.208

2.200 2.215

current (A)
2.3

22.4
27

30.7 37.8 38 43.2 48.9

50.0

484 823 957 1332 1952

1625

237

165

57
54

1.9

52.5 49 50 40 44
47

4.8 6.7 7.8 12.1 13.3

15.3

The slopes of both plots are essentially the same, and the enthalpy of activation for open-circuit hydrogen
evolution is calculated as ,-, 70 kJ mol-1. The similarity of slope is illustrated in Fig. 10, but it can be seen that there is an upward shift for the results of endof-life cells. This suggests that the mechanism for opencircuit hydrogen evolution is the same in both cases, but that the active area for evolution increases with age. This is presumably because of the progressive leaching out of antimony from the positive grid by corrosion, followed by its deposition on the negative active material. Hydrogen is preferentially evolved from local cells on the negative plate between the spongy lead and the deposited antimony [2]. The open-circuit, hydrogen-evolution rates were used to calculate the rate of self-discharge. The data are summarized in Fig. 11. At 20 C, the self-discharge rates of old and new cells are found to be 50 and 14% per year, respectively; this data should be c o m p a r e d with that of Linden [9]. In the present case, the results show an increased t e m p e r a t u r e dependency. It should be r e m e m b e r e d , however, that the rate is determined by many factors, including the length of storage, Tables 3 and 4 summarize float-charge data for old and new cells, respectively. T h e cells were maintained on float-charge, with no t e m p e r a t u r e compensation of the applied voltage. The float currents achieved for old Table 3
Hydrogen-evolution data for end-of-life cells on float-charge Temperature (C)
H 2 flow H2

Table 5 Individual hydrogen-evolution rates for old and new cells on floatcharge at 38 C and the individual end-of-discharge voltages (EDV) at the end of C/5 rate discharge of the battery

Cell number

Old cells
H2 flow

New cells

(SCCPHPC)
1 1731

EDV (V)
1.634

HE flow (SCCPHPC)
915

EDV (V)
1.662

2 3 4 5
6

1604 1659 1680 1664

1678

1.301 1.766 1.766 1.760

1.712

960 964 981 944

974

1.690 1.674 1.683 1.665

1.680

7 8 Mean
Standard deviation

1831 1463 1664

104.9

1.722 1.738 1.675

0.157

965 955 957

20.6

1.706 1.689 1.681

0.014

and new cells are similar, and there is a significant increase in the oxygen content of the evolved gas compared with the open-circuit case. This suggests that there is a Faradaic contribution to gas evolution. The hydrogen content of the evolved gas was greater for the old cells, but did not show any t e m p e r a t u r e dependence. For the new cells, there is some indication of a decrease in hydrogen content as the t e m p e r a t u r e increases. Figs. 12 and 13 again demonstrate a logarithmic relationship of hydrogen-evolution rate with temperature. In this case, the slopes are not the same, and the old cells demonstrate a greater t e m p e r a t u r e dependence. From the regression equation, the hydrogen evolution rate per Ah of rated capacity for new cells was found to range between 0.0166 and 0.241 cm 3 h - 1 Ah -1 at STP, at 20-50 C, respectively. For old cells, the corresponding rates were found to lie between 0.0298 and 0.903 cm 3 h -1 Ah -1 at STP. The results highlight the significant impact that cell t e m p e r a t u r e makes on the rate of hydrogen evolution during float charge, when the applied voltage is not

(SCCPHPC)

(%)

Mean cell voltage (v) 2.208

2.208

Float current (A) 1.2 2.6 4.3 5.5 7.2 10.0

2.3

19.5
23

198 519 855 1209 1664 5184

7658 398

64
69

25 29.5 34 38 43
51.5

69 67 63 65 64
71

2.214 2.215 2.211 2.210 2.211

2.208

16.3

J.B. Lakeman / Journal of Power Sources 53 (1995) 99-107

105

_c

4 0.00305

I 0.0031

I 0.00315

I 0.0032

I 0.00325

I 0.0033 1/T (ILK)

I 0.00335

I 0.0034

I 0.00345

I 0.0035

I 0.00355

Fig. 8. Plot of In H 2 flow r a t e ( S C C P H P C ) vs. T - t (K - t ) for end-of-life cells on open-circuit.

8.

7.

o c c

6.

4 y = -8826 l x 3 0.003 I 0.00305 I 0.0031 I 0.00315 I 0.0032 I 0.00325 1/T (ILK) I 0.0033 I 0.00335 I 0.0034 I 0.00345 I 0.0035

Fig. 9. Plot o f In H 2 flow r a t e ( S C C P H P C ) vs. T - t (K - t ) n e w cells o n open-circuit.

8.5

7.5

6.5

5.5

4.5

3.5 0.00305

I 0.0031

I 0.00315

I 0.0032

I 0.00325

I 0.0033 1[1" (l/K)

I 0.00335

I 0.0034

I 0.00345

I 0.0035

I 0.00355

Fig. 10. Plot o f In H 2 flow r a t e ( S C C P H P C ) vs. T - I (K -1) for end-of-life and n e w cells o n open-circuit.

compensated for temperature. For new cells, there is a 14-fold increase in rate on moving from 20 to 50 C; for old cells it is a 30-fold increase. In the confines of a submarine battery tank, individual cell temperatures can vary widely, depending on location. Therefore, large variations in gas-evolution rate can be expected, but

effective temperature compensation is not feasible. A similar situation will pertain in any large bank of batteries or cells and adequate precautions must be taken. The two batteries were discharge tested in order to establish the relationship between the discharge per-

106

ZB. Lakeman / Journal of Power Sources 53 (1995) 99-107 1

10

~ 100 "o

1000 5

I 10

I 15

I 20

I 25

I 30

I 35

I 40

I 45

I 50

I 55

Temperature / C Fig. 11. Plot of rate of self-discharge (% year -~) vs. temperature for end-of-life and new submarine cells.
10,

9.

v ~

7.

6,

5 0.003

I 0.0031

I 0.0032 1/T (l/K)

I 0.0033

I 0.0034

I 0.0035

Fig. 12. Plot of In H2 flow rate (SCCPHPC) vs. T -~ (K -1) for end-of-life cells on float-charge. 9, 8.5, 8, 7.5

6.5,

6, 5.5 5,
4.5

4 0.003

I 0.00305

I 0.0031

I 0.00315

I 0.0032

I 0.00325

I 0.0033

I 0.00335

I 0.0034

I 0.00345

! 0.0035

1/'r (l/K)
Fig. 13. Plot of In H2 flow rate (SCCPHPC) vs. T -~ (K -~) for new cells on float-charge.

formance and the hydrogen-evolution rate in cells. At the C/5 rate, the end-of-life battery of cells was found to have suffered a 13.7% reduction in capacity over its service life, whilst the battery of new cells achieved the rated capacity. Table 5 lists the individual hydrogenevolution rates for old and new cells on float-charge

at 38 C. The standard deviations indicate the increased variability in behaviour of old cells compared with new cells. End-of-discharge voltages (EDVs) are also listed for the individual cells at the end of a C/5 rate discharge. Again, an increased spread in performance is seen for end-of-life cells. The data also demonstrate that there

ZB. Lakeman / Journal o f Power Sources 53 (1995) 99-107

107

is no correlation between the discharge performance of individual cells and the corresponding hydrogenevolution rate. The same behaviour is also observed for cells on open-circuit. The individual cells are consistent, however, in their hydrogen-evolution rates and in their discharge performances. End-of-life cells were found to have suffered a significant reduction in discharge performance over their service life, but the rate of hydrogen evolution on float- or open-circuit significantly increases over the same period. This means that although the active area for electrical discharge has reduced, the active area for hydrogen evolution has increased, probably because of the progressive transference of antimony.

5. The open-circuit hydrogen evolution rate can be used as an indicator of the rate of self-discharge of the negative plate. 6. On open-circuit, the rate of self-discharge of the positive plate is significantly lower than that of the negative, and progressively falls with time. 7. Gas-evolution levels when float charging at ~ 2.2 V/cell are significantly higher than on open-circuit, and the oxygen content is higher. 8. Float charging without compensating the applied voltage for cell temperature can lead to a dramatic increase in the rate of hydrogen evolution as the temperature rises.

References

4. Conclusions 1. For cells of the same age, the gas-evolution rate from a cell on open-circuit, or on float-charge, does not correlate with its electrical discharge performance. 2. As a cell ages, its gas-evolution rate increases but its discharge performance decreases. The mechanism for open-circuit hydrogen evolution does not change with age. 3. The active area for hydrogen evolution increases with age, but the active area for discharge decreases. 4. Hydrogen evolution rates at open-circuit and floatcharge vary logarithmically with temperature.
[1] J.B. Lakeman, J. Power Sources, 27 (1989) 155-165. [21 N.E. Bagshaw, J. Power Sources, 40 (1992) 113-122. [31 M. Maja and N. Penazzi, J. Power Sources, 22 91988) 1-9. [41 M. Johnson, S.R. Ellis, N.A. Hampson, F. Wilkinson and M.C. Ball, J. Power Sources, 22 (1988) 11-20. 15l L.T. Lam, J.D. Douglas, R. Pillig and D.A.J. Rand, J. Power Sources, 48 (1994) 219-232. [6] S. Venugopalan, J. Power Sources, 48 (1994) 371-384. [7l H. Waterhouse and R. Willows, Br. Patent No. 622 512 (1949). [8l A.J. Salkind, G.E. Mayer and D. Linden, in D. Linden (ed.), Handbook o f Batteries and Fuel Cells, McGraw-Hill, New York,

1984, pp. 14-17.

[9] D. Linden, in D. Linden (ed.), Handbook o f Batteries and Fuel Cells, McGraw-Hill, New York, 1984, pp. 13-18.