Corrosion of ARMCO iron in HClO₄ solution in laixed water-alcohol solvents. I. Potential effect of solvent

A a n D H I Ï Ï E S I I A T I S L O n i E H S I S FOLIA CHIKICA 5, 1985

Henryk Scholl

CORROSION OF ARMCO IRON IN HCIO^ SOLUTION IN MIXED WATER-ALCOHOL SOLVENTS. I. POTENTIAL EFFECT OF SOLVENT

ARMCO iron anodio dissolution in HCIO. solutions in mixed water-alaohol solvents (MeOH, EtOH, PrOH-1,

PrOH-2, Etdiol, Prdiol-1,2) has been investigated by means of potentiodynaaic (v « 0.002 V method. The con­

tributions of the liquid junction potential (A and of the standard potentials scale shifts

(û® ) in the corrosion gotential shifts between wa­ ter and mixed solvente ( a ® V have been deter­

mined. w corr

The mechanism of electrochemical anodic dissolution of iron in aqueous HgSOj and HCIO^ solutions containing non spe­ cific adsorbed ions ha3 been described in a number of publi­ cations [1-7 3. It is generally agreed that hydroxyl ions play a major role in this process. Depending on the degree of energetic heterogeneity of the polycrystalline surface of the iron electrode, a catalyzed or noncatalyzed mechanism i3 pro­ posed. In the rate determining step of the electrode reaction, a decisive role is played by active FeOH’*' complexes,

The process of anodic dissolution of iron becomes more co­ mplex when it takes place in electrolyte solution in mixed so­ lvents consisting of water and nonaqueous solvent miscible with water* The complexity of the process results from the fo­ llowing effects»

1° Adsorption of the molecules of nonaqueous solvent on the electrode surface. The problem has received considerable attention of investigators, mainly in terms of the theory os. inhibitive action;

2° Adsorption of solvent molecules leads to a shift of ae­ ro charge potential of the metal CpQE 5 » 0)}

3° Molecules of the nonaqueoua solvent participate in the formation of active complexes» The meohaniam of electrode re­ action hao been proposed in this case for only a few mixed solvent«« water with DMSO [8, 9 j water with IMP Cto], and wa­ ter methanol iMeC?l)[10] and ethanol (EtOH)[11] j

4° Changes in the values of electrode potentials characte­ rising ';h® reaction are the sum of the interactions between solvent dipoles and the electrode (a ®1 and the change of the »tandard free enthalpy of ion 2 solvation during tra­ nsfer from one solvent to the other (a!!i _{Op r Jl}u f X The construe-ti an of the eleotrochemical cell may alio cause the value of the liquid junction potential difference (a^1 (f to con­ tribute to the measured potential difference.2

The chief ooncern of the present study was the last of the above mentioned problems, which comprises the set of effects commonly referred to as the solvent effeot in eleotrode reac­ tion.

Experimental

The ntudy of the electrochemical corrosion of AHHCO iron in HC104 solutions in mixed water-aliphatic alcohols solvents was conducted-using "Tacuasel" equipment (PHT 10.5 potentios- tat, ''Servovit" linear sweep generator, "Logalex" logarithmic amplifier, and "Sefraa" XY recorder)»

An Pe-AHiiCO electrode "Prolabo" 99*99 % in the form of wire 2.5 cm long and 0.05 cm in diameter was annealed in an

electric oven at 1 « 1220 K for T * 4 hrs in a hydrogen atmo­ sphere, cooled in hydrogen atmosphere, and polished eleotro- chemically in 1 M HCIO^ in dimethyl eter of ethylene glycol ("diglym") at 258 K. Following rinsing with water, the eleotro­

de was subjected to another course of the above thermal tre­ atment, It was then mounted in a Teflon holder and Inserted

in a tri-electrode measuring cell. The space of the working electrode was separated from the counter electrode space (ft) by means of porous glass. All measurements were conducted in solutions deaerated with argon (99.99 Ji) at 298 K.

The measuring celle were set according to the following schemest

(A) tPejO.5 M RCIO^ (.mixed solvent) | aqueous electrolytic bri-dge| S<3, (B) : ?e |0.5 M HC10. ¡electrolytic bridge) 10~2 U L1C1, AgCl, Ag

> ■ - * - ... ... . . v

in the mixed solvent

In the c e l l Ca) the classical electrolytic bridge(KNO^ t KC1 * 4 t l)was used» The meaouring cell and this bridge woe separated by the cell including the mixed solvent under Inves­ tigation.

Uee was mad« of mixed solvents of with ethanol, propanol-1 (rrOH-1), propanol-2 (PrOH~2) , «thandiol CKtdiol) , and pro- pond ioI-1,2 (Prdiol) • All alcohols, manufactured by "Carlo Kr- ba", were p.a, and the concentrations used ranged from 0 to 94 mole %.

The working electrode employed in cne of the measurements was annealed, polished electrolytically, subjected to cathodic polarization, brought to equilibrium potential, activated at that potential for two minutes, and then aubjected to cathodlc polarization again. The solution waa replaced pneumatically, and the measurement proper was conducted at the eweop rate of polarization potential of v ■ 0.002 V s~1 yielding polarization curve lg I « fCs). The above heat treatment procedure applied to the electrode ensured reproducibility of résulta within 3 %•

Results and discussion

The quasipotentiostatic curves 1g I ■ f(E)cbtained made it possible to determine the characteristics of the electrode pro­ cesses with hydrogen depolarization and activated dissolution of iron. The Tafel slopes Pflb+ and ?ilb_ , which characterize the anodic and cathodic processes respectively, as well the order of the hydrogen ion depolarization reaction arc collec­ ted in Table 1. As can be seen from those data, the Tafel coe­ fficient of the anodic dissolution reaction ia constant for all HCIO^ solutions in the mixed solvents. At the sa.ie tina, its - b value, equal to 40 + 5 mV/dec.. suggests that « H e

dissolution proceeds according to the catalyzed reaction me­ chanism [2-4,6],

?ig. 1, Potentiodynamio(v»1•10“3 V.s“1) polarization curves

of \RKC0 iron in 0.5 K }!C10, fiolutionst

la) a - H-O; b - 84 nol % EtOH; c - 84 mol % PrCK-2 j d - 72 mol % Etdiol; 1b) Tafel*s curves.

i/easuranento conducted in 0.5 H HCIO^ solutions in solvents with high alcohol content made it possible to obtain polari­ sation curves [Fig. 1],Similarities in the ourveo indicate that aioilar anodic iron dissolution mechanisms are involved. The a'cser.oe of a current "plateau" in the range of polariza­ tion potentials corresponding to passivation processes seams to suggest that anodic oxidation of alcohol taken place, :?or mocohydroxyl alcohcls undergo anodic oxidation more readily than diole.

Tab!« 1« Kinetios data of the Fe/Fe(Il)reaetion in UC104 (T » 298 K) In nixed water - alcohol solvents. ^

Solvent ..'W w b+ [a V ] h . O V] HgO 1.0 40 +2 120 +10 H20 - MeOH 0.9 ... 1.2 40 +5 120 +10 H20 - MtOH 0.9 ... 1.2 40 +5 120 +10 H20 - PrOH-1 0.9 ... 1.1 40 ¿5 120 -*10 HgO - PrOH-2 0.9 ... 1.1 40 ¿5 • 120 +10 _mm H 20 - Etdiol 0.8 ... 1.2 40 ¿10 120 +20 HgO - Prdiol 0.8 ... 1.4 40 +10 120 +20 10 1 - • Uerr [ A * C m *]

Pig. 2. Dependence i corr - f(X) for AKMCO iron in 0.5 M HC10, solution« In mixtures* a • HgO-Etdiol; h - HgO-Pr d 5 :

Flg. 3. Curv«s 8it0„r / Wicorr versus the solvent composition» « ) HgÖ-UeOH; ( - 0 - 0 - ) HgO-ßtQH; (--- )HgO--PrOH-2.

On the basis of the Tafel elopes values of the corrosion current density were obtained (for geometric surface). Pig. 2 illustrates these values in the form of the dependence

i __ «

f ( x )

ccrr s(

¿■iy. 3 depicts the dependence = f(X) for the mixed sol­ vents of water with monohydrcxylL alcohols.

Corronlon of ARMCO iron in RC10, solution ___ ________________ • ___ 4

The plots of Pig, 2 exhibit a minimum of 1_ _values for corr

madia containing ~ 20 mole * alcohol. It has to be kept in »ind here that this corresponds to the composition of cany cooling mixtures C 50/50 vol % )ccemonly employed in industrial equipment. The increased value of the corrosion current densi­ ty for solvents containing more than 20 mole % diol may be at­ tributed e.g. to the complexing effect of the organic compo­ nent C 12 3 . s.,

The * f(x) plots shown in Pig, 3 have shapes "corr

typical of electrodic reaction inhibited by electrically neut­ ral organio molecules [13-153, also confirmed in studies of iron corrosion [10,11,16] . In the case of email ROH concen­ trations in the solvant, inhibition of the anodic dissolution process takes place. In a wide range of alcohol concentrations in the solvent ( 1 0 + 90 mole 56), the magnitude of the corro­ sion current is limited by the change of structure of the bcI- vation complexes. On the other hand, for ROH concentrations of > 90 mole % the shape of the plot in Pig. 3 suggests the pos­ sibility of the inhibiting effect of water, a possibility also noted by other authors [16,17],

An impostant role in the interpretation of corrosion pro­ cesses is played by the value of the potential associated with the electrodic reaction, a cirsumctance that is frequently ig­ nored in corrosion studies conducted in mixed solvents. As a ruke the reaction potentials arc determined with reference to saturated calomel electrode (in aqueous solution) , and shifts in the values of the electrodic reaction potentials are then interpreted as being a property of the reacting system, negle­ cting the eum of the efects deriving from the eolvont [13-20] «

In addition to producing the effects listed in the intro­ ductory part of this paper, solvent composition aleo gives ri­ se to a change of the standard free enthalpy of hydrogen ion solvation during its transfer from one solvent to the other, an effect of particular significance in the case when standard hydrogen potential scale is employed. At the same time, solvent composition may also be responsible for the appearance of 1: <;u- id-liquid interface in the measuring cell. If the letter sffe-

liquid junction potential(a *1 y> appearing at the inter­

face. ■

An attempt to deal with the above problem« was made by B a c a r e l l a and S u t t o n [11J taking account of

the work of P o p o v y o h [ 2 1 ] , I s m a i l o v and A l e k s a n d r o v [22,24], and C r u n w a l d [25-27]on the "degenerate proton activity coefficient"*

For the measuring cell set in accordance with the scheme adopted Trom the paper [11]i

C O

!«•* « V ° 4 I SI I KC1eat.aq.' H«2012- “» where

*'t T °L m aL

E ‘° ♦ E? - E*

C1)

The magnitude of the corrosion process potential is given by equation!

EFe “ Ecell * EL + ESCE * ^2)

where, in turn, E®eil is an experimental value.

The magnitude of E^ was determined by measuring the po­ tential difference in the cell [11]i

( D ) Pt, H 2 ( H + , CEtOH-HOH)i|E®|| XClM t ^ f Hg2Cl2Hg .

The difference in the Galvani potentials proposed by B a c a r e l l a and S u t t o n [11] is equal to

$1? “ $ S C E " Ecell “ ( ^ M ” 1*8^ + ( t s ^ 1>s2 ) + (<i)82“ $SCE

where C — 4>s ) « is the difference in the Galvani potentials between a point on the metal(Fe or Pt) and sane point S^ in the solution phase, (cj>s - <J>S ) * - E® is

the potential difference at the liquid-liquid interface bet­ ween the mixed solvent and saturated aqueous solution of KC1 Ccf. eqn. 1), and (<£o - <^3CE') ESCE difference

in the Oalvani potentials for tha reference electrode relative to aqueous KC1 solution.

Measurement of the potential difference in cell ( D ) leads to the deterainatlon of acidity as a measure of proton aotivi- ty in water at a given point in timet

pA - - lg aH+ — 1* . yj}* . fH+ f ( 4)

where yjj+ is the conventional molar lyonium ion activity co­ efficient in solvent S, and fg+ is the "degenerate proton acti­ vity coefficient" measured for infinitely dilute solution rela­ tive to water as the reference state [11].

Since, at the same time, for aqueous solutions

- lg ajj+ - - lg cH+ • yj5+ - pH , ( 5 )

Bacare11a and Sutton in effect cbtained*

E* . 0 .0 6 lg (cj^ • y{j+ • % * • > CEcell(:pt)+ESCB) C 6 )

In the discussion that followed immediately after the pu­ blication of the study under consideration, Schwabe [28] cate­ gorically dismissed any possibility of determining the absolu­ te scale of acidity under the conditions described by tho au­ thors [11]» Accepting Schwabe’s objection, we can state the following* the quant, ty E^ (eqn. 1), which B a c a r e l -1 a and S u t t o n £11] term "liquid Junction potential" comprises the liquid Junction potential difference pi'oper and the diffusion potential at the KGi8a*.aq. K H 2^°4 ilJterfa-ce. It can be easily found, by the method of succa3ive appro­ ximations and making use of Planck’s theory [29] , that the di­ ffusion potential at that interface is some milivolts and may be practically neglected ;

- the magnitude of the liquid Junction potential differen­ ce at the 0.5 H (Et0H-H0H)f0.5 M H,.504 (H0H) interface

cannot be accurately determined directly from SEH measurement in the proposed cell ( C ). There are, however, a number of in­ direct procedures for determining the value of A ® f ^ in the function of solution composition [21,30,31 J • Calculations

can be approproately bastid only on *ery similar and independe­ ntly obtained a ® (pl m f(X)valuesi

- the degenerate proton aotivity coefficient determined by I z m a i l o v [22-25] and G r u n w a l d [25-27] la, in effect, a measure of changes in the otandard free enthalpy of hydrogen ion solvation during its transfer from the mixed or nonaqueous solvent to water a ® y jr+# • At the same time, it io a measure of the shift of the standard hydrogen potential scale. Using different methods, I s m a i l o v [22-250, S t r e h l o v [ 3 2 ] , and P a r s o n s [33 J,aa well as the present author D O ,31]have all determined the value of a J y. for water-alcohol media and obtained satisfactory agre­ ement of the results.

Pig. 4 shows the plots of A ® <¡pT ■ f (X)(curve 1) and A ® HH +* » f(XXcurve 2) obtained for the solventa used in the present study;

- in the left-hand branch of cell ( C ) as proposed by Bao- «ralla and Sutton, the value of the reference is 3 p ^+® . The value cf SgCE in the right-hand branch is also referred

to arbitrary aero of the hydrogen scale, i.e. aleo to W ^ijj+*» Summing up the above, the "liquid junction potential* of U a c & r e l l a and S u t t o n [11.] is the sum of di­

ffusion potential, liquid-liquid potential difference

-'■* * i n and the potential difference resulting form chan­ ge cf the standare free enthalpy of hydrogen ion solvation. The values of E? obtained by Bacarella and Sutton and thooe determined in our own research [30,3 1 ] for the sum

A A fa » f(X) are listed in Pig. 5.

Pig. 6 illustrates the values of the differences in the sxperinontally obtained corrosion potentials plotted in the function of solvent composition. The A ® ®corr *a tbe figure applies to the value obtained from measurement in the cell containing liquid-liquid interface (a), while A ® ECorr io for the result obtained from measurement in cell (B). In the latter cell, the calculation of the silver chloride sil­ ver electrode potentials in the mixed solvents in each caue relative to the hydrogen scale zero point adopted for each solvent vras based on the literature valueB of ¿g activl j coefficients [34-37]. Taking account of the sum

Pig* 4. Potential changes curves» 1 - A ® 9 i m * 2 - A j ^ n+* « fix* A ® gdlp - f(*> tor mixed solvents (— — — — } HgO - KeOH»(' - )HgO-EtCil;

Pig. 5. Comparison of the data« 1 - E® (Bacarella’e and Sut­ ton’s [11] and 2 -(A ® + A ® ) author*«

data [30,31]

£ 8 5?. s i ® J*H+ ) for each of the mixed solventa, we can propose equations yielding one standard potential scale ^ w ^ccrr^ for ■BeaBur'e<* corroalon potential differen­ ces.

? I ¿ _{w corr} 9 I. * ¿ * _w y corr + ¿ î h + A í _{' *}f / >

A 8 E w corr

1 " ? L A i Ecorr(Me0H> 2 ' A w Ecorr(MeOH^

3 - y T A ® 2 „ t * - Bacarella'a and Sutton's da-

w corr _

ta [11]; A-thlawork; 4 - A w Scorr(EtOH);

y T A® E _{li w corr} (P rO H -2) ; 6 - i “ _w E _corr (P r O H -2 );_f

The values of A f <p _ calculated on the basis of the pro- w T corr

posed equations (7 ) and (8 ) in the function of solvent composition are collected in Pig. 7.

The quantity g|ip is the potential effect of the intera­ ctions between solvent dipoles and the electrode aetal. The concept has been extensively discussed b y J a k a s s e w - s k i [38], T r a a a t t i £39],and B o c k r i a [40]

Pig. 7. The values of A ® 9 corr calculated fora equations (7) - and (8) -according to tho coapoaitions of the solvents* (--- ■} H«0 - KeOH; ( ---)g j?0 _ EtoajC--- ) H20 - PrOH-2; g ?0 - Etdi&l .

for metallic eleotroda in zero charge potential,

Th« value of aero charge potential for iron FeE 5 - 0 in aqueous solutions has not been definitely established, and for HgS04 solutions it ranges from -0 ,3 to -0 .3 7

[41,42]. At the sane time, Kelly claims that F«‘'corr < Fe^'S » ■ 0, The values of Fe^S ■ 0 in mixed solvent« or nonaque- oue solutions are not available, which makes crodible inter­ pretation of the contribution of g^ip to the determined va­ lue A ® ^corr vei*y difficult. However, a comparison of the 8i

- - - - f(X)plot <?ig. 3)with the aJ g}lp - f(X)plot

*eorr

0

(Pig. 4) and

Aw

9 corr * f(XXPig« 7) seams to indicate

This paper was supported by the framework MR-I-11 Problem

References 1. P. H i l b e r t , Y , i i i y o s h i , G. E i c h k o r n , ff. J. L o r e n z, «T. Electrochea, Soc., 118. 1919 (1978) 2. K.P, B o n h o e f f e r , Z. H e u s l e r , Z. Pfcys. Chen. N.P., 8 , 390 (1956); S. H e u s 1 e r, Z. f. Elektrochcm., 62, 582 (1961)

3. J.O'M, B o c k r i s , A.K.ii. R e d d y , "Modern Electro­ chemistry" vol. 2., Plenum Rosetta, NY 1973

4. J. O'li B o c k r i s , D. D r a z i c . A . D e s p i c , Electrochim. Acta 4 , 325 (1961); J. 0 *M. B o c k r i s , H, K i t a, J, Electrochem. Soc,, 108. b7£>(1961); J. O'M. B o c k r i a , D, D r a z i c, Electrochim. Acta. ?, 293 (1961)

5. I.H. P 1 o n s k i, J. Electrochea, Soc., VMS, 146 (1969); ibid ¿16, 1688, (1969); ibid m , 944 (1970)

6. E,J.

K e

7. Zynter Y.D., A. L, P. a t i n i a n, Slektrohkira., 2 S 1371 (1966)

H. Soboli 8 * H, S c h o 1 1, B. G r u l i k,. Elactrochi*. Acta 2 749 (1978) 9. K. S c h w a b a, Z. f. PhyB. Chamia Jf.P., Bd 108. 61 (1977) 10. C. A* F a r i n a, G. F a 1 t a, P. 0 1 i ▼ a n i , Cor- roaion Soi., 1£ 465 (1978) 11# A. i*. B a o a r a l l a, A* L* S u t t o n, J* Bltotroohui« Soc., 122, 11 (1975)} ibid 122, 1636 (1975) 12. L . E , T 0 y g a n k o v a , V. I* V i g d o r o v i t c h , Zastoh. Met., 1^, 436 (1977) 13. W. J a a n i o k a , P. H. S o h w a i t a a r , Z* Phyak. Cham., ¿2, 104 (1967) 14. J. L i p k ó w . a k i , J. Elaotroanal. Cham., ¿2, 333 (1979) 15. J. L i p k o w 0 k i, Z. 0 a 1 u a., J. Elactroanal. Chaa.

48, 337 (1973); ibid 61, 11 (1975); ibid ¿8 , 91 (1979) • 16. R. P. T o b i a a, K e n N o b a, J. Elaotrocham. Soc., 122. 65 (1965) 17. V. X. V i g d o r o v i t o h , L. E. T a y g a n k o v a, S. V. 0 « i p o T », T. K o r n e a v a, Zastoh. Mat., 594 (1973) 18» A. C a r q u a t t i , F. M a z z a , Corroaion Sci., 1JJ, 337 (1973) 19. B. R. G a b a, Corroaion Sci., 175 (1973) 20. V. I. V i g d o r o v i t c h , L. E. T e y g a n k o v a , N. 7. ? i 1 i p o t a, Zastoh. Mat., 10, 427 (1974) 21. 0. P o p o y y c h , A . J . D i l l , Anal. Chan., ¿1_, 456

(1969) 0. P o p o v y c h , Crit. Rav. Anal. Cham., 73, ( 1970)

22. 5. A. I z m a i 1 o v, Zh. Piz. Khim., 1142 (1960^ "Elektrochimia Raatvorov", Uoscow (1976)

23. V. A. A 1 e k s a n d r o v, N. A. I z m a i 1 o v, Zh. Pi*. Khim., ¿1. 2619 1957; ibid ¿2, 401 (1958) 24. N. A, I z a a i 1 o v ., Zh. Piz. Khim., ¿J., 1142, (1960) 25. B. G u t b e z a h l , E * G r u n w a l d , J. Am. Cham. Sec., 22, 565 (1963) 26. 2. G r u n w a l d , B. J. B a r k o w i t z , J. Am. Cham. Soc., 22» 4939 (1951)

27. E. G r u n w a l d , S. W i n a t a i n , J. Am. Cham. Soc. 20, 846 (1948)

28» K. S c h w a b « , J. Eleotrocham. Soc., 122. 1636 (1975) 29. W* B* M o r f , Anal* Chea*, £ £ 810 (1977) 30. B* J a k u s z « w a k i f H. S c h o l l , Electrochim* Acta 12, 1105 (1972) 31* B. J a k u a s f l w a k l , li. P r z a s n y s k i , U* S o h o 1 1, A . S i e k o w a k a , Electrochia* Acta 2 2 i 119 (1975) 32. H. M. K o • p p, H. W • n d t, H. S t r e h 1 o w, 2. El- ektrochem*, fa£, 483 (i960)

33* B. C a a «, R* F a r a o n a * , Trana Faraday Soc,, 63. 1224,(1967)

34* G* J a n z, R* P. T, T o « k 1 n a, "Nonaqueoua Bleotro- lyt« Handbook", Acad. Praaa, NY 1973

35. D. J* G. I v a a, G. J a n z , "Rafarencaa Electrodes" - Th«ory Practico", Acad. Pr«aa, NY 1961

36* N* R a b l n d r a , W. V e r n o n , J. G i b b o n s , A. L. B o t h « « l l , J * Elactroanal, Chea., 101 (1972) 3 7 * N . R a b i n d r a R o y , W* V • r n o n, A* L* B o t h - w • 1 1 , J, Electrocham. Soc., 118. 1302 (1971) 38* B* J a k u a z • w a k i, J. Cham* Phys*, ¿1., 846 (1959) 39* S. T r a a a t t i , J. Elactroanal* Cham*, 351 (1971); ibid 19 (1974); ibid ¿4, 437 (1974) 40. J. O'M. B o o k r i s, M* A* H a b i b, J. Electroanal* Cham*, 68, 367 (1976)

41* E* 0. A i a z i a n, Doki. Akad. Nauk. SSSR 100. 473(1955)

Institute of Chemi­ stry, University of Łódź, 90-136 Łódi,

Henryk Scholl

KOROZJA ZELAZA ARMCO W ROZTWORACH HC104 W MIESZANYCH ROZPUSZ­ CZA ŁNIKACH WODA-ALKOHOL. I. POTENCJAŁOWY EFEKT ROZPUSZCZALNIKA

* Na podstawie ąuasipotenojoatatycznych (v ■ 0.002 V*s”1) krzywych polaryzacji wyznaczono parametry reakoji anodowego roztwarzania żelaza ARMCO w roztworach HCIO^ w wieszanych ro­ zpuszczalnikach woda-alkohol C MeOH, EtOH, Pr(JH-1, PrOH-2, Et- diol, Prdiol-1,2 ) , Do obliczenia wartości przesunięcia poten­ cjału korozji ( A ® ^corr^ prty Prz«J®oiu od wody *o miesza­ nego rozpuszczalnika zaproponowano równanie uwzględniające przesunięcie skali potencjałów ( A ® i różnice poten­ cjałów na granicy faz ciecz-ciecz ( A ® ).