Thermodynamics of the unsymmetrical mixed electrolyte hydrogen

Thermodynamics of the Unsymmetrical Mixed Electrolyte HCI-LaCi,. Rablndra N. Roy, James J. Gibbons,. Department of Chemlstry, Drury College, Sprlngfie...
0 downloads 0 Views 570KB Size
J. Phys. Chem. 1983, 87, 2365-2369

2365

Thermodynamics of the Unsymmetrical Mixed Electrolyte HCI-LaCi, Rablndra N. Roy, James J. Gibbons, Department of Chemlstry, Drury College, SprlngfieU, Mlssourl 65802

J. Chrlstopher Pelper, and Kenneth S. Pltrer' Department Of Chemlstry and La wrence Serk8ley Laboratory, University of Calltorn&, Berkeley, Callfornia 94720 (Recelved: December 2, 1982)

The contribution of higher order electrostatic terms (beyond the Debye-Huckel approximation) has been investigated for the system H+-La3+4--H20. Emf measurements were carried out on solutions at temperatures from 288.15 to 318.15 K, and at ionic strengths from 0.1 to 5.0 mol kg-', by using hydrogen and silver-silver chloride electrodes. From these measurements it is possible to determine the Pitzer mixing coefficients and +H,La,CI as a function of temperature. They are well represented by linear equations yielding temperature-invariant estimates of aSBH,La/aT and a+~,L~,a/aT. These estimates may be used to predict the relative apparent molal enthalpy for aqueous solutions containing H+, La3+,and C1- with reasonable confidence. A brief table of the enthalpy and the activity coefficients is included.

Introduction At finite concentration the properties of an electrolyte solution are complex functions of short-range forces between solute ions and of long-range (electrostatic) forces. The theory of Debye and Huckel assumes only the latter, together with a uniform hard-core repulsion, and leads to a knowledge of the limiting behavior of such solutions. "Higher order" limiting laws arising from purely electrostatic effects may also be developed from a more complete theory; Friedman' discusses these, including a law for unsymmetrical mixing of ions of the same sign. Since short-range interionic forces have substantial effect on solution properties at rather low concentration, it is not easy to isolate the purely electrostatic terms in treating experimental measurements. It was shown by Pitzer2 that the limiting law for unsymmetrical mixing could be extended to an expression valid at finite concentration and that the inclusion of this term resolved an anomaly in the properties of the system HCl-AlCl,. The importance of this electrostatic function for unsymmetrical mixing was demonstrated impressively by Harvie and Weare3 in their calculations on mineral solubility in the system Na+-K+-Mg2+-Ca2+-C1--S042--H20 at 25 "C. The present study extends our understanding of unsymmetrical mixing effects by very precise measurements over a range of temperature on the system HC1-LaC13. The measurements are treated in the system of equationsH for electrolyte properties based on a Debye-Huckel term plus virial coefficients (or interaction coefficients) for short-range interionic interactions. Pair and triplet interactions usually suffice, but the second virial coefficients are functions of ionic strength in an electrolyte. The higher order electrostatic function for unsymmetrical mixing is easily included as a contribution to the second virial coefficient. The temperature derivative of the mixing parameters for the Gibbs energy is related to the heat of mixing, and (1) Friedman,H. L. 'Ionic Solution Theory"; Interscience: New York, 1962. (2) Pitzer, K. S. J. Solution Chem. 1975, 4,249. (3) Harvie, C. E.; Weare, J. H. Geochim. Cosmochim. Acta 1980,44, 981. (4) Pitzer, K. S.J.Phys. Chem. 1973, 77, 268. (5) Pitzer, K. S.;Kim, J. J. J.Am. Chem. SOC.1974, 96,5701. (6) Pitzer, K.S. In "Activity Coefficients in Electrolyte Solutions"; pytkowicz, R. M., Ed.;CRC Press: Boca Raton, FL, 1979; Vol. 1, Chapter 7. Peiper, J. C.; Pitzer, K. S. J. Chem. Thermodyn.,in press.

in an accompanying paper' the experimental data on the heat of unsymmetrical mixing are treated. That paper reviews the basic theory and the related thermodynamics; hence, those aspects will not be repeated here.

Experimental Section Potentials for the cell

Pt,H2(g)lHCl(ml),LaC13(m2)IAgC1,Ag (1) were measured at 288.15, 298.15,308.15, and 318.15 K, and at ionic strengths from 0.1 to 5.0 mol kg-'. Solutions were prepared from reagent-grade hydrochloric acid, which was twice distilled to azeotropic composition, and reagent-grade lanthanum chloride, which was recrystallized 3 times from aqueous solution. Analyses of stock solutions of these reagents were carried out gravimetrically by precipitation of silver chloride. The molalities of HCI and LaC1, were determined to within &0.01% and *0.02%, respectively. Sample solutions were prepared by mass dilution of the stock solutions. Electrode p r e p a r a t i ~ n ,purification ~.~ of hydrogen gas, cell design, temperature control and stability (k0.02 K), and other experimental details have been given elsewhere.lOJ1 Measurements of the emf were made with a Leeds and Northrup type K-5 potentiometer in conjunction with a Leeds and Northup Model 9829 dc null detector. Stable, drift-free readings were always obtained, even in the most concentrated solutions. Equilibrium was assumed to have been attained when the emf remained constant (to well within 0.1 mV) for 30 min at ionic strengths below 1.0 mol kg-', and for 15-20 min above 1.0 mol kg-'. All data are listed in Table I. As a check on technique, standard potentials of all AgAgCl electrodes were determined at all experimental temperatures. Cell 1 was used with m l= 0.01 mol kg-' and m2 = 0 mol kg-', as suggested by Bates et a1.I2 The standard potentials thus obtained were in excellent (7) Pitzer, K. S.J. Phys. Chem., preceding article in this issue. (8)Bates, R. G. "Determination of pH", 2nd ed.; Wiley: New York, 1973; p 33:. (9) Robinson, R.A.; Roy, R. N.; Bates, R. G. J.Solution Chem. 1974, 3, 837. (10) Roy, R. N.; Gibbons, J. J.; Trower, J. K.; Lee, G. A. J. Solution Chem. 1980,9, 535. 1965, No. 271, 28. (11) Bates, R. G. NBS Tech. Note (U.S.) (12) Bates, R. G.; Guggenheim, E. A.; Harned, H. S.; Ives, D. J. G.; Janz, G. J.; Monk, C. B.; Prue, J. E.; Robinson, R. A.; Stokes, R. H.; Wynne-Jones, W. F. K. J . Chem. Phys. 1956, 25, 361; 1957, 26, 222.

0022-365418312087-2365~0 1.5010 0 1983 American Chemical Society

2366

The Journal of Physical Chemistty, Vol. 87,No. 13, 1983

Roy et ai.

TABLE I : Experimental Potentials for the Cell Pt,H, IHCl(m, ),LaCl,(m,)IAgCl,A@ I 0.10 mol kg" 1 0 3 m /( mol kg-' ) 100.00 69.559 50.020 29.831 lo%,/( mol kg-' ) 0 5.073 8.330 11.695 E(288 K, 95.72 kPa)/mV 367.14 378.42 395.07 E(298 K, 95.64 kPa)/mV 351.85 365.93 377.74 394.90 E(308 K, 95.67 kPa)/mV 363.69 375.81 393.61 E ( 318 K, 95.64 kPa)/mV 361.07 373.43 391.64 10% ,/(mol kg-') 103m,/( mol kg-' )

E(288 K,95.67 kPa)/mV E ( 298 K, 95.63 kPa)/mV E ( 308 K, 95.67 kPa)/mV E(318 K, 95.63 kPa)/mV

232.17

103m,/(molkg-') 103m,/(mol kg-') E ( 298 K , 95.64 kPa)/mV

2000.0 0 185.08

10'111 I /(mol kg- ' ) 10' m,/( mol kg- ' ) E(288 K, 95.76 kPa)/mV

3000.0 0

E(298 K, 95.66 kPa)/mV E(308 K, 95.76 kPa)/mV E(318 K, 95.76 kPa)/mV

150.64

lo%,/( mol kg-' )

4000.0 0

10 'm,/(mol kg-' ) E(288 K, 95.38 kPa)/mV E(298 K, 95.37 kPa)/mV E(308 K, 95.38 kPa)/mV E(318 K, 95.38 kPa)/mV 103m,/(molkg-' ) 103m,/( mol kg- ' ) E( 288 K, 95.34 kPa)/mV E(298 K, 95.34 kPa)/mV E(308 K, 95.29 kPa)/mV E(318 K, 95.30 kPa)/mV a

1000.0 0

120.70

5000.0 0 93.44

I = 1.0 mol 702.43 49.595 252.15 247.34 242.39 236.38

19.904 13.349 407.23 407.47 406.54 404.97

9.805 15.032 427.32 428.16 427.95 427.14

kg-' 494.96 84.174 265.67 261.38 256.38 250.99

293.03 117.83 283.47 279.77 275.46 270.63

187.974 135.34 297.14 294.20 290.00 285.51

106.98 148.84 313.89 311.48 307.97 304.18

I = 2.0 mol kg-' 1371.7 991.63 104.72 168.06 203.99 218.32

585.02 235.83 238.95

386.84 268.86 253.45

190.39 301.60 275.32

I = 3.0 mol kg-' 2079.5 1466.0 153.42 255.66 177.65 195.56 171.65 189.82 164.84 183.46 157.46 176.67

898.01 350.33 215.59 210.22 204.46 198.08

597.15 400.48 230.70 225.87 220.75 214.89

317.67 447.06 250.05 246.12 241.14 235.69

I 4.0 mol kg-' 2827.2 2047.1 195.47 325.49 151.02 168.00 144.30 161.60 155.26 129.86 147.54

1269.5 455.09 189.86 184.15 178.30 170.97

839.74 526.71 206.06 200.86 195.09 188.45

384.33 602.61 231.99 227.70 222.60 216.73

I 5.0 mol kg-' 3435.8 2354.2 260.70 440.97 127.68 150.22 120.62 143.68 113.97 137.15 106.48 129.85

2092.4 484.60 155.67 148.99 143.03 135.35

1580.4 569.93 169.20 163.31 156.73 149.78

1106.1 648.89 184.53 178.91 172.85 166.30

I is the ionic strength.

agreement with previously published ~a1ues.l~

Theory and Equations The Nernst equation for cell 1 is E = Eo(Ag,AgC1)-

reference, we reproduce those equations here. The activity coefficient of HC1 in a solution containing LaC1, is given by (YHC1)

= f'

+ ( m H + mCl)(BH,Cl + mClCH,Cl) +

+ %ICLa,CI + OH,La) + mHmCl(B'H,Cl + + mLamCl(B$a,C1 + CLaCl + 1/ZJ/H,LA,CI) +

mLa(BLa,CI cH,Cl)

mHmLa(e'H,La

Herein, Eo(Ag,AgC1) is the standard potential for the silver-silver chloride electrode, F is the Faraday constant, and R and T retain their usual meanings. The ionic activities are given by the product miyi, where m has dimensions of mol kg-'. The hydrogen fugacity may be assumed equal to the hydrogen pressure P(H2) in this work. The quantities m, and Poare included to ensure dimensional correctness and have the values 1.0 mol kg-l and 1atm = 0.101325 MPa, respectively. In light of the results of Bates et a1.,12 concerning the behavior of silver-silver halide electrodes, values for E" (Ag,AgCl) were taken from ref 10. All physical constants were taken from ref 14. The activity coefficient of HC1 appearing in eq 2 is given for a mixed electrolyte by Pitzer and Kims (their eq 15) and more recently by Pitzer2 (his eq 26-30). For ease of (13) Roy, R. N.; Gibbons, J. J.; Bliss, D. P.; Baker, B. K.; Casebolt, R. G. J. Solution Chem. 1980, 9,12. (14)Cohen, E. R.; Taylor, B. N. J.Phys. Chem. Ref. Data 1973,2,663.

p

= -A+[

-k %ZJ/H,La,Cl) (3)

1

2 + - In (1 + bI1l2) 1+ b P 2 b 1lj2

(3a)

Herein, p is the Debye-Huckel function for the activity coefficient with parameter A,. Values of A,, with dimensions of kg1I2 are given by Bradley and Pitzer.15 The parameter b has the standard value 1.2 kg1/2 The molality of ion i is given by mi and the ionic strength (15) Bradley, D. J.; Pitzer, K. S. J. Phys. Chem. 1979,83, 1599.

The Journal of Physical Chemistry, Vol. 87, No. 13, 1983

Thermodynamics of the Mixed Electrolyte HCI-LaCI,

by I; both have dimensions of mol kg-’. The second and third virial coefficients for a pure electrolyte ij are B, and Cij and have dimensions of kg mol-’ and kg2 mol-‘.i, respectively. The second virial is assumed to have some dependence upon the ionic strength, as represented above with an equation in two adjustable parameters, pi!) and pi!’. These parameters have the same dimensions as Bi .; a has the value 2.0 kg’l2 B f j is the ionic strengtk derivative of Bidwith dimensions of kg2 mol-2. The ionic charges, in protonic units, enter into the formula for the third virial as zi and zj and relate Cij to the parameter Cp. which was originally defined for the osmotic coefficient ana is normally tabulated. The mixing coefficients eH,La and $HhcI arise from the difference in interaction between H+ and La3+ from the appropriate mean for H+-H+ and La3+-La3+and are defined in several Using the notation of Pitzer,2 one may decompose OH,L~ into electrostatic and short-range contributions ‘ d H h and %H,La. The term EdH,&a has an ionic strength derivative EL9‘~h;the corresponding term for short-range forces StJ~h is assumed to be independent of I. The electrostatic terms and may be calculated a priori from formulas given by Pitzer.25 zizj (44 Eei,j = z [ J ( x i j ) - f/ZJ(xi,i)- l/zJ(xjj)l

(4b) where zi, z. and I have been defined above. The univariant function (and its derivative, J’) is obtained from cluster-integral theory with omission of short-range forces. The variable xij is defined by

f

xij

- ~z,z~A+P/~

2367

TABLE 11: Ion-Interaction Parameters at 298.15 K

H C1

LaCl,

0.1775a 0.2945a C@ 0.0 00 80a 104(a p ( O ) / a 2’) -3.08 1 1 0 ~ ( a p ( ~ ) / a ~ ) 1.419‘ 104(ac@/aT) 0.6213c

p(0)

p(1)



a

From ref 16.

‘ From ref

17.

0.5889’ 5.60‘ -0.02383’ 2. 5267b 79.80’ -3.7144’

‘ From ref 18.

these integrals; the authors prefer a 200-point Gauss-Legendre method. and Ee’HJaderives The temperature dependence of solely from that of A,, as indicated in eq 5. This knowledge, plus precise emf measurements interpreted via eq 2 and 3, allows the determination of ‘ ~ H , Land ~ IC/H,La,C1 as functions of temperature. This analysis assumes that the virial coefficients appearing in eq 3 are known over the temperature range of interest. This is in fact the case if it may be assumed that, e.g. Pt2,dT) = P&1(298.15 K)

+ (T - 298.15 K)(d~~2,ci/dT)zss.i5~

and similarly for the remaining pure-electrolyte parameters. Table I1 gives values for all HC1 and LaC13 parameters, and their temperature derivatives, at 298.15 K. Thus, only the mixing parameters %Hsg and $ H h ~ l remain unknown. The excess enthalpy of HC1-LaC1, mixtures may be calculated from the excess Gibbs energy6p7by the usual differentiation. For these mixtures Hex is given by

(5)

The functions J and J‘ are given by Pitzer2i7in integral and series form. The latter is useful only for xij I0.10, which, fortuitously, is the region where numerical integration is difficult. Dropping for the moment the subscripts on x , one obtains J(x) = -C (In x p=3

+ Kp)xP-1pP-3/(p!(p - 3)!)

(6)

The quantities K are constants given as sums involving Euler’s constant, p , and integral fractions; the first few values are K3 = 0.41970, K4= -0.292608, K5= -0.369464, and K6 = -3.00381. From eq 6 one may obtain

k

m

J’(x) = -E [l + ( p - l)(ln x

+ Kp)]xP-2pP-3/(p!(p - 3)!)

p=3

(7) The error in truncating either series at p = 6 for x = 0.10 is less than 0.570, which was calculated by comparison to the direct numerical result. For x > 0.10, numerical methods are most precise. The integrals given by Pitzer2p7may be rearranged via variable substitution to yield

J(x) = X - 1+ 4

J’(x) =

Here, fH is the Debye-Huckel function for the enthalpy, with parameter A H (of dimensions J kg1/2moF3l2)as defined by Bradley and Pitzer.15 Once more the mixing parameter a9H,La/dTmay be separated into electrostatic and short-range effects

lly(l

y4 + J1(t/q2)[1

- eq) dt

- (1- q)eq] dt

q = xt/ln t

The electrostatic contribution may be calculated from eq 4a

a

-(Eei,)

dT

=

From eq 5 one obtains

(84 (8b) (8c)

A variety of numerical techniques may be used to compute

Results and Discussion Linear least-squares procedures were used to fit eq 2 and 3 to the experimental data. Only isothermal fits were performed. All points were given equal weight as mea-

2368

The Journal of Physical Chemistry, Vol. 87, No. 13, 1983

Roy et al.

TABLE 111: Mixing Parameters from Isothermal Fits t o Data of Table I 1 5 "C

25 "C

35 " C

45 "C

0.296 i 0.010 -0.019 i 0.005 0.37

0.319 i 0.015 -0.045 i 0.007 0.62

-0.296 i 0.021 0.084 t 0.010 0.86

-0.284 i 0.030 0.059 i 0.015 1.15

With Ee and E e ' Included H,La/(kg mol-' ) @ H.La,Cl/(kg2 0 f i t ImV

0.265 c 0.013 0.030 c 0.006 0.49

0 H,La/(kg mol-' )

-0.305 i 0.022 0.129 i 0.010 0.71

'0

0.278 0.009 0.41

Without H.La,Cl/(kg2 mol-2) ofit/mV

?: ?:

0.009 0.004

and Ee'

-0.324 i 0.017 0.118 i 0.008 0.72

TABLE IV: Activity Coefficients and Excess Enthalpies for the System H+-La3+-Cl--H20 at Ionic Strength Fraction y of LaCl,"

0.010 0.020 0.050 0.100 0.200 0.400 0.600 0.800 1.000 2.000 3.000 4.000 5.000

0.905 0.875 0.830 0.795 0.766 0.755 0.765 0.785 0.811 1.011 1.318 1.761 2.390

0.903 0.874 0.828 0.793 0.762 0.747 0.751 0.765 0.784 0.937 1.171 1.500 1.952

0.901 0.871 0.824 0.788 0.756 0.736 0.735 0.743 0.756 0.867 1.039 1.275 1.590

0.715 0.641 0.538 0.465 0.406 0.367 0.358 0.361 0.369 0.466 0.652 0.968 1.497

0.724 0.652 0.551 0.478 0.416 0.373 0.358 0.354 0.357 0.414 0.533 0.7 28 1.034

0.732 0.661 0.560 0.487 0.423 0.375 0.355 0.346 0.342 0.366 0.433 0.544 0.7 09

0.001 0.002 0.008 0.023 0.063 0.173 0.314 0.479 0.666 1.845 3.295 4.866 6.428

0.001 0.002 0.008 0.023 0.061 0.162 0.289 0.435 0.600 1.685 3.238 5.341 8.100

0.001 0.002 0.008 0.023 0.062 0.168 0.302 0.459 0.637 1.811 3.454 5.593 8.279

" The trace activity coefficients of HC1 and LaCl, are given in columns 4 and 5, respectively. The excess enthalpy is given for a quantity of solution containing 1.0 kg of water. 0.35

with € 8 , €8' A without Ee'

Ee,

-I-

T 0 05

-1'

I

0

I

I

I

2 3 (mH+mC,)/2, mol kg-'

I

I

I

I

t

I

4

Figure 1. The upper plot shows the compositlon dependence of (A In Y w ) / m , at 298 K, eq 13. Data for ( m , imc,)/2 < 0.2 mol kg-' are not shown: the uncertainties for these points are large and the deviations are negative and very iarge if and are neglected (the triangles). The straight lines represent least-squares fits to the emf data. The lower plot shows the residuals (E - Ecalcd)for the data at 298 K, for least-squares calculations with and without higher order electrostatic terms, circles and triangles, respectively. Note the definite dish-shaped curvature for the fit without these terms. Obviously, the data at low Ionic strength are much better represented with inclusion of and

surements of E. The pure-electrolyte parameters were taken from ref 16-18. For each isothermal fit, the only unknown parameters were SOH,La and $H,La,CI. The unknowns s O ~ , band $H,J&l may also be estimated by a graphical procedure. This procedure defines the quantity A In yHclas the difference between the experimental value of In Y H C ~ ,calculated from 2 eq, and that (16) Pitzer, K. S.; Mayorga, G. J. Phys. Chem. 1973, 77, 2300. (17) Pitzer, K. S.;Peterson, J. R.; Silvester, L. F. J. Solution Chem. 1978, 7, 45. (18)Silvester, L. F.; Pitzer, K. S. J . Solution Chem. 1978, 7, 327.

-0.05 I

I

I

I

15

25

35

45

T/"C

Figure 2. Temperature dependence of the mixing parameters sBH,La The fitted lines are given by eq 14. and

calculated from eq 3 with This yields

%H,La

and

$H,La,C1

equal zero.

SO that a plot of (A In y H c J / m h VS. '/Z(mH + mcl) should give a straight line with intercept s9H,4, and slope Figure 1 shows at the top a plot of this type for the data at 298.15 K. Care must be exercised when using this procedure, since a constant uncertainty in the measured emf leads to an uncertainty in (A In yHcl)/mLawhich is greatly magnified for small La3+molalities. However, we recommend construction of such plots as a visual check on the least-squares calculations. The quality of the measurements and of the fit is better judged from the deviation of measured from calculated values of E shown in the lower part of Figure 1. When the EO and terms are included (solid circles), the deviations are randomly positive and negative and rarely exceed 0.5 mV.

J. Phys. Chem. 1983, 87, 2369-2373

Also plotted in Figure 1 are the results (as triangles) if the higher order electrostatic terms are omitted. The systematic departures are apparent, primarily for dilute solutions. The values are given in Table I11 of 'OH,L~, # H , L ~ , c ~ ,and the standard deviation of fit determined at each temperature via the least-squares procedure, and Figure 2 presents the temperature dependence of the results graphically. The error bars correspond to the standard deviations of the parameters as determined by the fitting procedure. Also included in Table I11 are values of the standard deviation and of OH,Le and #HhLa,C1determined without E O H , h and EO'Hb; inclusion of the higher order electrostatic terms is clearly indicated by the standard deviation values as well as by Figure 1. The mixing parameters are well represented by the equations = 0.281 + 0.0018(T - 298.15) #H,b,CI

= 0.006 - 0.0025(T - 298.15)

(14)

The temperature coefficients of these parameters are then

a

-('OH,La)

aT

a dT

-(#H,La,C1)

= 0.0018 f 0.0007 = -0.0025 f 0.0003

(15)

where, again, errors were determined via the least-squares fitting procedure. Most of the earlier measurements's21 for this system are (19) Randall, M.; Breckenridge, G . F. J. Am. Chem. SOC.1927, 49, 1435.

2389

not sufficiently precise to contribute to those aspects considered here. The most recent study, that of Khoo et al.,n reports data with precision comparable to the present study but only for 25 "C and only to I = 3 mol kg-'. They used slightly different pure-electrolyte parameters and there are small differences in the reported E values for pure HCl. In general, the results and conclusions of Khoo et al. for 25 "C are the same as those of this investigation. The activity coefficientsof HC1 and LaC13and the excess enthalpy were calculated by eq 3 and 9-11 and are presented in Table IV. The enthalpy of mixing is the difference between the next to last column (y = 0.5) and the average of the adjoining columns (y = 0, 1). These calculations are reasonably straighforward; hence, we have not made efforts to provide extensive tables.

Acknowledgment. Acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, for partial support of this work, and to the Director, Office of Energy Research, Office of Basic Energy Sciences, Division of Engineering, Mathematics, and Geosciences of the US.Department of Energy under Contract no. DE-AC03-76SF00098for additional support. We thank Kennan Buechter and Steven Faszholz for their assistance in the electrochemical cell measurements. Registry No. HCl, 7647-01-0;Lac&, 10099-58-8. (20) Lietzke, M. H.; Stoughton, R. W. J. Phys. Chem. 1967, 71, 662. (21) Lietzke, M. H.; Danford, M. D. J. Chem. Eng. Data 1972,17,459. (22) Khoo, K. H.; Lim, T.-K., Chan, C. Y. J.Solution Chem. 1981,10, 683.

Influence of the Physical Structure of Irradiated Starches on Their Electron Spin Resonance Spectra Kinetics Jacques J. Raffl' and Jean-Pierre L. Agnei Service de Radio-egronomie, CEN Cadarache, 13 115 Saint-Paul lez Durance, France (Received: February 19, 1982; In Final Form: September 9, 1982)

This study deals with the shape and kinetic changes of the ESR spectra of eight irradiated starches, from several hours to several months after y-irradiation. Whatever the origin and water content of the starches two major radicals or groups of radicals are observed. The kinetic law depends on the water content; two main zones are pointed out which are relative to the amorphous and crystalline parts of starches.

Introduction The chemical and toxicological studies that we carried out on y-radiolysis of starches'$ naturally lead w to inquire into the mechanism of radiolysis of these starches. Electron spin resonance (ESR) studies have been undertaken for a long time on the "polysaccharide radicals" induced by y radiation or chemical initiation. As the observed phenomena are very complex, the most recent works moved toward the use of low temperature^,^,^ hydrogen (1) J. Raffi, L. Saint LBbe, and G . Berger, Food Preseru. Irradiat., Proc. Int. Symp., 1977, 1, 517 (1978). (2) J. Raffi, J. P. Agnel, C. ThiCry, C. FrCjaville, and L. Saint-LBbe, J. Agric. Food Chem., 29, 1227 (1981), and references cited therein. (3) G . Abaghian and A. Apresyan, Arm. Khim. Zh., 32, 850 (1979).

donor^,^ or sensitizers.6 These methods allow the study of the initial process (excited states, radical ions, primary radicals) but they yield only little information about the chemical structure of the long-lived radicals, i.e., those remaining after months of storage at room temperature; however, such radicals are the more interesting from a toxicological point of view,'P2 in order to ascertain the wholesomeness of irradiated food. Moreover the results (4) A. Plonka, J. Bednarek, and H. Zegota, Z. Naturforsch, B, 34,1525 (1979). (5) G . Abaghian, J.Magn. Reson. Relat. Phenom., R o c . Congr. AMPERE 20th, 173 (1979). (6) A. Merlin and J. P. Fouassier, Angew. Makromol. Chem., 86, 123 (1980).

0022-365418312Q87-2369$Ql.5010 0 1983 American Chemical Society