The system graphite-MnCl2-AlCl3 Kinetics, structure and mechanism of formation

Curbon Vol. 22. No. I, pp. 5341. Prtnted m Great Britain

WOM223184 $3.00 t 00 Pergamon Prev Ltd

1984

THE SYSTEM GRAPHITE-MnClrAlC13: KINETICS, STRUCTURE AND MECHANISM OF FORMATION T. DZIEMIANOwICZ,t and W. FORSMAN Department of Chemical Engineering, University of Pennsylvania. PA 19104,U.S.A. and R. VANGELISTIand A. HEROLD Laboratorie de Chimie Minerale Appliquee, University de Nancy. France (Receioed

25 January

1983)

Abstract-Graphite-MnCll-AK& intercalation compounds have been preparedvia MnClz-AI& complexes. First stage compounds are eventually formed at both 325 and 500°C;products formed at the lower reaction temperature are richer in Mn (C~.~M~CI*(AICI,)~.~I) than those at 500” (C7,l~MnCl2(AICl2 &I,). At 325” the mechanism is a quasi-selective intercalation of AiCI~to the second stage followed by insertion of MnC12to a mixed stage I compound, and finally an isostage Mn enrichment with Al depletion. At 500”.insertion proceeds differently. Rate limitingreactions at each temperature are proposed. Powder X-ray diffractograms give Z, = 9.51a for Mn-rich first stage compounds. Comparison of calculated and observed (001) structure factors gives an Mn-Cl spacing of 1.46A as in the pure dichloride. Analysis of (h/d) positions and the (101) and (Ill)_ intensities of the inserted MnCI: indicate an intercalant superlattice with a0 = 3.69A and co = 3 x I, = 28.53A. First stage compounds prepared at 500”are considerably less ordered. Results obtained at 325”are especially significant in that (I) this represents the lowest temperature at which dichlorides have been intercalated to rich stages, and (2) long-range (i.e. 3-dimensional) ordering has been documented in a dichloride intercalation compound for the first time, and is sensitive to reaction temperature.

due to complexation:

1. INTRODUCTION

The insertion of transition metal dichlorides into the graphite lattice was heralded by Croft[l] in 1956. In the intervening quarter century, the list has grown considerably. A recent paper lists the chlorides of Cd, Co, Cu, Hg, Mn, Ni, Pd, and Zn as those known to intercalate in the presence of free chlorine[2]; FeCl* graphite intercalation compounds (GIG’s) have also been prepared, but only by reduction of FeCI, compounds[3]. Because of their exceedingly low vapor pressures-typically less than 1 Torr at 500°C-high reaction temperatures are required and rates of formation are slow. It appears, in fact, that this particular balance of thermodynamic and kinetic factors often precludes formation of rich (i.e. first stage) compounds. That is, at temperatures sufficient to give a measurable reaction rate, the intercalation desorption equilibrium (eqn 1) may be shifted to the left, allowing only formation of

MCI,(s) t nA1Cl,(g)~MAI,CI~,+~(~).

(2)

In this manner, non-volatile chlorides may be transported via vapor phase complexes. The increase in concentration of the metal, M, in the vapor phase is striking; the ratio of complex vapor pressure to that of the pure chloride at the same temperature may exceed lo’“! The results of Stumpp’s work are reproduced graphically in Fig. 1. As the quantity of AU, is increased its partial

A&!

MO nC + MCI,(s) t +(g)&MCl,+,(s)

(1)

dilute stage compounds. Nickel chloride, e.g. is inserted only to the second stage, and the chlorides of mercury, palladium and zinc do not progress beyond stage three. In 1977 Stumpp reported that the addition of AU, to transition metal dichlorides (CoCI,, NiCI,) or lanthanide trichlorides arguments their rate of intercalation[4]. This effect was attributed to a wellknown enhanced “apparent volatility” of the chloride

./Present address: Hercules. Inc.. 800 Greenbank Rd.. Wilmington. DE 19X08.U.S.A.

TIME (HRI Fig. I. Kinetic curves as a function of AICI, concentration for the system graphite-Co&.AICI1 as reported by Stumpp[4]. Numbers on curves refer to mg. of AICI3 reactant per gram of graphite.

T. DZIEMIANOWICZ et al.

54

pressure and that of the complex must both rise (at 500°C all of the AlC& will be in the vapor phase): P ~.41nc13n+z

=

Kn[~,a,l”.

(3)

It can be seen in Fig. 1 that the reaction rate increases accordingly. But this interesting work raised more new questions than it solved. For example, can intercalation via AlCl, complexes proceed at reduced temperatures, and if so, how low, and at what rate? And is it possible to produce richer compounds by this route than by direct combination of graphite and the dichloride? In an earlier paper[5], we presented preliminary results on the system graphite-MnClz-AlC& which may be summarized as follows: (a) Synthesis. First stage compounds were prepared from natural graphite powders at temperatures as low as 325°C; the richest compound prepared under these conditions had a composition by chemical analysis of

two-zone furnace such that the graphite temperature always exceeded that of the chloride mixture (AT = 25SO’C).On exiting the furnace, the reagent portion of the tube was quenched first to minimize the risk of condensing intercalant onto the graphite. Products were weighed in a glove box under dry air to determine weight uptake (Am/m,,), and samples were analyzed for carbon, hydrogen, chlorine and metal content by the C.N.R.S Service Central d’Analyse, Vernaison, France. Powder X-ray spectra were obtained by diffractometer (MO KOLradiation) in transmission with the sample sealed in a Mylar cellule. Pyrographite (001) and (hk0) scans were performed in similar fashion with samples sealed in glass capillaries. 3. RESULTSANDDISCUSSION

2.EXPERIMENTAL

3.1 Equilibrium products, kinetics and mechanism Kinetic data (weight uptake vs reaction time) are plotted in Fig. 2 for compounds prepared from graphite powder at various reaction temperatures. The agreement between measured weight uptake and that calculated from analytical data (as (1 - %C)/%C) is quite good. At both 325 and 5Oo”C,rich products are eventually formed which appear to be pure first stage compounds according to their X-ray (001) reflections. Surprisingly, reactions at 400” failed to yield a first stage compound. After 93 hr, a mixture of stages 3 and 2 (Am/m,, = 80%) is formed, and at 209 hr a second stage compound is obtained (Am/m,,= 160%) with only traces of the first stage. This observation is consistent with thermodynamic properties of the MnC&AICl, system. Mass spectrometry data of Binnewies[ lo] show that substantial amounts of MnAl& and MnAlCl, exist in the vapor phase at 325 and 500°C respectively, but that neither complex is as abundant at 400”. Thus it is not unreasonable that the reaction at 400 is a bit sluggish. Unfortunately, the data are insufficient to determine whether a first stage compound appears thermodynamically forbidden at 400” due to a low complex partial pressure, or if this merely poses a kinetic limitation.

Graphite-MnC&AlCl, compounds were prepared by the classic tube a deux boules technique. Ceylon graphite (particle size 540 p) and in some cases thin pyrographite plaques (- 10 x 2 x 0.2 mm) were placed in one end of the tube. The reagent portion of the tube contained anhydrous MnC& (as received) and anhydrous Al& (resublimed); the amount of A1C13was measured so as to give a total initial pressure of about 2atm. according to the equilibrium [7]

(a) Reaction at 325” Figure 3 shows the evolution of the X-ray powder diffraction spectra of graphite-MnC&-AlCl, compounds prepared at 325” with the (001), i.e. staging lines indicated. In conjunction with the analytical data presented in Fig. 4, the following pattern emerges. The graphite is first rapidly and selectively intercalated to stage 2 with Al& The ideal stoichiometry of this reaction,

2AlCl,(s)~AA1~C1,&)~2AlCl,(g)

18C(s) t I/2Al$&(g)F?C,~AICl~(s)

C~.9~MnCl*(AIC130.*,.

(b) Mechanism. At 325”, Al& is first selectively intercalated to stage 2, after which MnCl* enters the lattice. At 500” a different mechanism prevails. (c) Structure. Powder diffractograms indicate that the MnC12is well-ordered in the 325” compound and retains the rhombohedral (ABC) stacking sequence which it possesses in the pure state. In a related study[2], it was shown that the (hk0) reflections are identical to those of an MnClz compound prepared via the direct synthesis, with the carbon and MnClz a-axes at a 30” angle to one another. In the present study, we present further data on this system, the interpretation of which permit a better understanding of the optimum reaction conditions, mechanism of insertion and structure of the equilibrium products. Where possible, results are compared with the very recent report of Baron et al.[6] on MnCl* GIG’s prepared via the direct synthesis.

(5)

log,,K(mm) = -6749/T - 2.013 log,,T + 16.628. (4) For kinetic experiments, MnClz was always present in excess of that required to produce a first stage compound. Molar ratios of AlC13:MnClz were thus on the order of unity. It should be noted that physical mixing of the graphite with intercalant was scrupulously avoided in order to avoid subsequent washing of the products, a common technique with other investigators. Reaction tubes were sealed under 0.5 atm. CIZ and placed in a

compares favorably with the analytical result (C,6.4AIMn0.3C13.n). This is followed by a regime during which Mn content increases at roughly constant Al content (Fig. 4) accompanied by a change in stage from 2 to 1. Finally, Mn content rises to its equilibrium value, but with Al depletion in a process which is iso-stage. It is noteworthy that at the longest reaction times (25 d.). the total metal chloride content determined by chemical analysis exceeds the liiniting compositions C4.5MnC12and

The system

graphite-MnClrAICI~:

kinetics,

structure

and mechanism

I

of formation

/I----L

8

4 TIME tilt Fig. 2. Kinetic

curves as a function

of graphite

temperature

poq = 0.5 atm. (0.0,

(“C) for the system graphite-MnClz-AK&:

A measured;

0, I,

pont~,, -- 2atm

A analysis).

44 AlC$ ‘ 4 Fig. 3. Time

evolution

of powder

t

/

I

8” 0 IMo KaJ

diffractograms for MnClrAICI~ GIG’s included for comparison.

C,AICI, for the richest pure component intercalation compounds.+ There is, however, some evidence of adW~AICI~ is commonly observed stoichiometry for saturated first stage aluminum chloride compounds. One arrives at a composition CqsMnCl? by coGparing the hexqonal lattice parameters, a~. for graphite (2.4 A) and Mn~l~~3.48 A). The ratio of the areas of the hexagonal cells (projected perpendicular to the c-axis) is thus (3.68/2.46)‘= 2.24. Since there are two carbon atoms per cell versus a single MnC12 unit, this ratio must be doubled. thereby giving C/MnClz = 4.5.

/

12O prepared

at Z?“C.

Pure Ai&

compound

sorbed MnC& in the X-ray spectra. It is conceivable that after the intercalation compound becomes “saturated” with intercalant, MnCI, continues to be transported via the vapor phase complexes, but finding no further vacancies within the lattice is forced to remain outside, Le. on the graphite surface. Figure 5 depicts schematically one possible mechanism-based on the folded sheet model of Daumas and Herold[ll]--which is consistent with these results. After the formation of rhe stage 2 compound (a). MnCI~

56

T. DZIEMIANOWICZ et al.

.”-.a3 -%

lMd1

-02 I

L

I

1

4

TIME Fig. 4. Mn(O) and AI(O) contents of MnClrAU

A.

- AlC13

vuvo

“&

(D.1

GIG’s prepared at 325°Cas a function of reaction time.

is gradually added to the lattice-initially filling the interlaminar spaces to give a mixed first stage compound (b). This step is consistent with the increase in Mn content-which is manifest in both the analytical data (Fig. 4) and the growth of the MnCl, (110) peak (Fig. 3)-at constant Al level. As the proportion of Mn rises further, it can do so only at the expense of aluminum. This entails an exchange of the metal chlorides, probably mediated by MnAl,Cls.+2 molecules at the edges of the graphite planes. Unlike the previous step, this process occurs at constant stage (no change in (001) positions in the upper traces of Fig. 3). With some MnC12 now in

0

I

8

every interlaminar space, consolidation and long range ordering of the MnClz domains begins (c), signaled by the appearance of the (101) and (111) complexes on the X-ray pattern. Finally, only small islands of AK& remain, either kinetically trapped at the interior of an intercalant layer, or perhaps ionically bound as AlCL-either of these is a plausible explanation as to why it seems impossible to remove all of the aluminum from the lattice under these conditions, even at long reaction times. Stepwise mechanisms such as this are not unheard of in the intercalation literature. Freeman [ 121 added

@ - MnCl,

/ ,

0”

000 o o

o

-12.8i

Fig. 5. Mechanism of MnClrAICI~insertion at 325°C.(a) Formation of stage 2 AK12compound. (b) bi-insertion to give mixed first stage compound, (c) further Mn enrichment with onset of long-range order, (d) first stage MnCl2compound with AlClj impurities.

The systemgraphite-MnClz-AlCb:kinetics. structureand mechanismof formation N205 to a dilute graphite-ferric chloride intercalation compound and obtained a ternary of reported composition C3,FeClJNz05), ,. Presumably the NzOs is inserted in the vacant interlaminar spaces, though no X-ray data are presented. Thallium halides will also undergo sequential insertion reactions. Stumpp[l3] has shown that graphite-thallium bromide compounds which have a limiting composition of stage 2, may be reacted with TICI, to yield a mixed first stage compound C12.5TlC11.8 Br, + The most directly analogous reaction, however, is that of the formation of the ternary donor compound KH&. According to Lagrange et al.[14], insertion proceeds as follows. When pyrographite is exposed to the saturated vapor of a KHg amalgam, potassium is selectively inserted first, progressing through weaker stages until stage I KC, is obtained (it is pointed out that this step is not completely selective: some mercury is soluble in the intercalated potassium sheets, just as there is some MnC12in the initial stage 2 AICI, compound in the present work). In the second step, mercury and more potassium are inserted cooperatively to give the final product KHgG with its characteristic K-Hg-K trilayered intercalant. As in the MnClz reaction, this is an iso-stage process, though the interlayer spacing expands markedly from 5.23 A in KCs to 10.16A in KHgC.,. AICI, and MnClz layers, on the other hand, have nearly identical values of I, -about 9.5 A. Further support for a sequential mechanism is provided by the following experiment. If, rather than using an excess of MnCIZ, the reactant mixture contains only enough of the dichloride to yield a second stage compound (calculated as C,MnCI,), the results are as follows. At equilibrium, the MnC12 appears to be completely consumed, X-ray (001) lines indicate a first stage compound (1, = 9.5 A), and chemical analysis gives C ,2.,MnClz(AlClzh)n.ss. This corresponds well with a situation such as (b) in Fig. 5. All interlaminar spaces are filled. hut with no further MnClz available, the mechanism is arrested at the mixed first stage.

Stage evolution and compositional changes at 500°C are presented in Pigs. 6 and 7 respectively. It is clear that insertion proceeds somewhat differently at higher temperature. Once again a stage 2 compound is formed at the outset. but unlike the reaction at 325”, significant quantities of manganese are present even at short reaction times. At 12hr the manganese content is already comparable to that of the aluminum. and the (100) peak of the inserted MnClz is quite pronounced (compare the I2 hr trace in Fig. 6 with that at 10 hr in Fig. 3). Further enhancement of Mn content is quite slow; it is not until about 8 days that a first stage compound is finally observed. This is in contrast to the 325” case in which the Mn content rise\ more or less continuously after early formation of the stage 2 AICI, system. These results are readily explained in terms of rate limiting reactions at each temperature. As mentioned previously, at 325” the rate of MnCl? insertion is governed by reaction via the gaseous complex MnAI,CI,:

57

325”: 18Ct 1/2A12Cl6(g)~C,sAICI3(s) fast

(6)

Cr8AlC13(s)t 3MnAl&(g)ti 3C6MnC12(AICl,),,(s) t 3.2AI,Cl&) SIOM’. (7) At 500”, the vapor phase consists of no fewer than 6 species. In approximate order of decreasing vapor pressures, these are A12C16 - Cl? - AICh 9 MnAlClr b MnAl+& B MnCl2. The early appearance of manganese in the reaction products probably reflects fas-

ter kinetics of reaction (7) at 500”. However, since the partial pressure of the MnAl,Cl, complex is lower at this temperature the equilibrium is shifted to the left, and the product of that reaction is less rich in Mn than at 325”. Further insertion of MnClz is subsequently limited by the rate of reaction with the MnAICls complex-or by the faint pressure of MnCl,: 500”: 18C t 1/2Al+Zl,(g)~C,~AICl~(s) fast

(6)

ClsAICl,(s) t l.2MnA12Cls(g)“? l.2C,sMnC12(AIClJo~ t 2.8AlC1&) fart

(8)

C,5MnCl&41Cl~)05t MnAlCl&)* 2C75MnClz(AlCl,)03(s)+ 0.9AlClJg) slog

(9)

or C,sMnCl,(AlC13),, t MnC&(g)+ 2C,.,MnCl,AlCl,),,(s) t 0.25AIClJs) SIOM’. (10) 3.2 Structure of graphite-MnC12-AICY, compounds 3.2.1 Ceylon powder. (a) c-axis ordering. First stage compounds gave values of 9.51 ‘4 for I,., the c-axis repeat distance. This is identical to that reported by Stumpp and

I-

8”

12O

0 IMo Kal Fig. 6. Time evolution of powder diffractogram for MnC12-AICI, GIG’s prepared at SWC. Pure AK13 compound included for

58

T. DZIEMIANOWICZ et al.

1

06

.l6bt$-

A A

Fig. 7. Mn(0) and AI(O) contents of MnClrAlCI3 GIG’s prepared at 500°Cas a function of reaction time.

Werner[8] in the earliest study of an MnClz GIC, but slightly larger than found recently by Baron, et al.[6]. It must be remembered, however, that the initial presence of AIC13in the lattice could have had some bearing on the latter structure. Analysis of the (001) reflections was carried out as follows. Stoichiometry was first adjusted to take into account adsorbed MnC&. If ratios C: AlCI, = 9 and C : MnC& = 4.5 are assumed, and all AlC& is taken to be present inside the lattice (condensation is unlikely at 325’) then maximum filling of the graphite occurs at a composition of CdMnC12)o.s,(AICla)o.z,with about 9% adsorbed material. With the stoichiometric coefficients ni thus determined, the single remaining unknown is z*-the dimensionless distance between Mn and Cl layersand B, the Debye temperature factor

Fool = exp (- B sin28/hz)~ nifi cos (2nlzr) I

=

exp (- B sin2~/h*){0.21fAIt 0.91fi+,”t 6fe cos rrl t fe,(0.63 cos (0.294 nl) t 1.82cos 2aIz*)}.

The mined

AI-Cl

distance

by Leung

was assumed et al.[9]

(11)

to be 1.40 A as deter-

(alternatively,

one could treat all Cl atoms as identical, separated by a single distance from the plane of Mn and Al atoms; in this work, results for both approaches were identical). Simultaneous solution for z* and B was accomplished by minimizing the residue R=c

Table

I. Observed and calculated (001) structure factors MnClrAICIr intercalation compounds (Ic = 9.51 A)

(002) 001

I

obs

F obs

F talc

589

11.17

12.98

002

1000

-38.18

003

711

41.13

34.72

004

190

15.03

14.93

005

not

in

-38.15

observed

006

240

29.42

36.69

007

41

6.02

14.05

008

161

27.80

17.22

(b) Long-mnge order. The abundance of (hkl) reflections in the powder diffractograms of compounds prepared at 325” (top of Fig. 3) suggests long-range-i.e. 3-dimensional ordering-of the MnC12 layers. In pure MnCI,, Cl-Mn-Cl trilayers are stacked in a rhombohedral A-B-C sequence. The corresponding structural parameters are a0 = 3.86 A and co= 17.47A, with the atomic positions in the cell given by Mn: 000; rh Cl: OOu,OOU; rh

(13)

where u -0.25 for many substances with the CdC& structure[H]. The structure factors for (101) and (111) reflections calculated from

IlFoo~l - IFnoi,obrll (12) x I&m,obsI .

The minimum value of R is 0.20 and occurs at z* = 0.153 with B = 2.9 A’; calculated and observed structure factors for these conditions are listed in Table 1. This value of z* corresponds to an Mn-Cl distance of 1.4 A-about the same as observed in pure MnCl*, and only slightly greater than that reported by Baron et aJ.[6]. The small number of reflections and possibility of interference with (hkl) peaks probably account for the relatively high R value.

FhL,= c f,, exp {2ni(hx, + ky, + Iz,)} n

(14)

are listed in Table 2 with observed structure factors obtained by diffractometer wigh MO .Ka radiation and after correction by Lorentz-polarization factors. After intercalation the MnCl* (1 lo! reflection gives a slightly lengthened value of a0 - 3.69 A. If the MnClz superlattice has retained its A-B-C stacking on insertion, the c-axis parameter will have increased to co = 3 x I, = 28.53 A. Furthermore, the parameter u eqn 13) which enters into (hkl) structure factor calculations, would be changed

The system graphite-MnCl~-AICI~:kinetics, structure and mechanism of formation Tahie 2. Observed and calculated (ekes structure factors in pure MnCl: (a” = 3.68A. cn = 17.47A, ri = 0.25)

d,,,(f)

(hkl)

101

3.15

104

2.58

107

1.97

110

I .a$

113

1.76

116, 202

1.56

119

1.34

F

obs +

F

oak+

3fMFl

3fMn+ 6fC1 3fMn

3fM”+

6fCl

3fMn 3fMn- 6fCl 3fM"

47

50

100

100

40

44

85

91

38

59

-13

-21

31

32

* In arbitrary units

from 0.25 to 0.283, assuming an Mn-Cl distance of 1.46A as deduced in the last section. Hence, both the positions und magnitudes of the Fhrl would change in a predictable way on insertion. Observed (hkt) positions in Mnrich stage I compounds coincide exactly with cafculated values of dhkt using cn = 28.53 A. Table 3 lists calculated and experimental structure factors for the (lo/) and (11I) reflections. As predicted, the (107) peak has grown to surpass the magnitude of the (104) (compare with Table 2 for pure MnCI,). And in the (Ill) reflections, although observed magnitudes are slightly higher than calculated, the ratios FIlo: F1,3:Flth: F,,, are in close agreement, and fall off less sharply than in pure MnCl,. Using the method of Stokes and Wilson[l6], the breadth of the MnC& (110)reflection gives a crystallite size of 120A in the n-axis direction for the intercalated MnCI, lattice; this echoes the findings of Baron et ai. in their study of the direct synthesis of MnC12 compounds (Ceylon graphite was also used in that study). It is interesting that Mn-rich first stage compounds prepared at 500” yield Table 3. Ohserved and calculated ihki) structure factors in intercalated MnCb i d

(f:>

)

d&;f)

F

talc

+

F

ohs

101

3.18

3fM,- 1.22fC,

34

*

104

2.94

3fNn+ 4.11fCl

88

91

107

2.55

?fEl"+ 5.94fC1

100

100

II0

1.85

3fNln+ 6.00fC,

85

104

113

1.8:

3fR"+

3.45fC,

65

89

116

1.72

3fM". 2.03fC,

24

44

119, 021. 20:

1.60

3fMn- 5.79fC,

27

34

’ in

* arbitrary

* Coincides

units

with intense

(OO!) reflection

+

59

much less well-defined (sky) maxima; structural disorder at elevated reaction temperatures has also been reported for the direct synthesis[61. 3.2.2 Results with HOPG. Characterization of pyrographite products was less definitive than that of powders, Due to small sample size, chemical analyses were not feasible and weight uptake measurements would have been accompanied by a high degree of uncertainty. Diffractometer scans of (001) and (hk0) reflections did, however, provide some information. At 325”,the mechanism begins as in powders; at short times the (001) positions and intensities are those of stage 2 AICI, with no discernible MnCI, (110) reflection. After several days a stage 1 compound is formed, with some stage 2 remaining. Unlike powder samples, however, the (001) intensities are intermediate between those of a first stage AICI, and MnClz (for pure M&I, compounds, the (002) is most intense, whereas in AK& it is the (003)). Furthermore, there are a number (at least five for 0 < @< 20”; MO Kcu) of very weak peaks which cannot be indexed in terms of either stage 1 (I, - 9.5 A) or stage 2 (Ic - 12.85A). These peaks are, however, compatible with intermediate bi-insertion compound in Fig. Sb, with I, 19h Thus, while the power diffractograms are not su~ciently sensitive to confirm this step in the mechanism, HOPG X-ray data to lend some support (Fig. 8). The (hk0) reflections (Fig. 9) include both a strong MnClz (110) and weak reflections at lower angles which are identical to those found in pure AICI, GIC’s[l7]. Thus, even at long reaction times, the AICI, concentration in pyrographite products is su~cient to alter (001) intensities and to produce domains with in-plane ordering. The breadth of the sharptest AICI, (hk0) reflection (at - 4 A) indicates that the AK], domains are about half the size of those for MnCIZ--i.e. on the order of 50 A. 4. CONCLUSIONS

First stage compounds graphite-Mn~l*-Af~l~ have been prepared at 325 and 500°C via vapor phase complexes This synthetic technique allows formation of rich stage metal dichloride compounds at the lowest temperature yet reported. Furthermore. transport of the chloride via vapor phase complexes obviates the need to physically mix the reactants-a technique which requires subsequent washing of reaction products prior to characterization (a procedure which may leach intercalant from the edges of the interlaminar spaces[l8]). Baron et a/.[61 cite “filling coefficients” of 80-85% for first stage metal chloride compounds after an HCI wash. Results presented herein indicate that apparent fitting coefficients of unity have been achieved in 32.5”compounds If the MnC& does indeed form islands as proposed for other dichlorides[& 191,it is possible that the mobile AICls molecules fill the interstices between islands, thus creating a very densly packed intercalant layer. The mechanism of insertion at 325”is a quasi-selective intercalation of AK], to stage 2. followed by MnCIz addition to a mixed first stage and then iso-stage MnClz enrichment. The reaction at 500” proceeds differently, probably reflecting changes in the concentrations of the

T. DZIEMIANOWICZ et al.

60

I

I

5”

I

6”

I

12” 8

I

13”

IMoKd

Fig. 8. (001)diff~cto~ram of HOPG-MnCl~~Cl3 compound prepaied at 325°C showing evidence of b&insertion compound with 1, - 19A.

Fig. 9. (hk0) diffractogram of HOPG-MnClz-AK13 compound showing evidence of in-plane ordering of MnC12and AK& domains. various species present in the vapor phase. Rate limiting reactions at both temperatures have been proposed to

explain differences in kinetic and compositional data. Three-dimensional ordering of the intercalant takes place at the lower reaction temperature but is less evident at 500°C. The MnC12retains its A-B-C stacking on insertion, giving rise to a hexagonal lattice with parameters aa = 3.69 A and co = 3 x I, = 28.53 A. Positions of numerous (hki) lines support this structural hypothesis, as do structure factor calculations for the (101) and (111) reflections of the intercalated MnC12 lattice. This is the only report of 3-dimensional ordering in a dichloride GIC and again is attributed to the lowered reaction temperatures which are only feasible through ~omp~exation, Finally, this work sheds new light on the merits of Stumpp’s initial report of intercalation via AK& com-

Production of rich stages at greatly reduced temperatures is a significant practical advantage of this reaction pathway; formation of well-ordered compounds is an additional bonus for the intercalation scientist, and makes this technique an attractive one for future studies of dichloride intercalant systems.

plexes.

Acknowledgements-This research was supported by NSF INT 79-16390and Army Research Office grant DAAG 29-79-C~208.

REFERENCES

R. Croft, Aust. 1. Chem. 9. 184(1956). 2. R. Vangelisti and A. Herold. Ext. Abs. Proc., 15th Bienn. Cant: Carbon, 381 (1981). 3. B. Liengme,M. Barlett, J. Hooley and J. Sams, Phps. tett. 25A, 127(1967). 4. E. Stumpp, Mat. Sci. &gng 31. 53 (1977). 1.

The system

graphite-M&12-AICh:

kinetics,

5. T. Dziemianowicz. R. Vangelisti, W. Forsman and A. Herold. Ext. Abs. Proc.. 15th Bienn. Conf. Carbon. 379 (1981). 6. F. Baron, S. Flandrois, C. Hauw and J. Gaultier, Solid State Commun. 42. 759 (1982). 7. E. Dewing, Metal. Trans. 1, 2169 (1970). 8. E. Stumpp and F. Werner, Carbon 4, 538 (1966). 9. S. Leung. M. Dresselhaus. C. Underhill. T. Krapchev, G. Dresselhaus and B. Wuensch. Phys. Rer. B24, 3505 (1982). IO. M. Binnewies..Z. Anorg. AII8. Gem. 437. 25 (1977). II. N. Daumas and A. Herold. C. R. Acad. Sci. Paris, Ser. C 268. 373 (1969). I?. A. Greeman, J.C.S. Chem. Commun.. 746 (1974).

structure

and mechanism

13. E. Stumpp,

Carbon

16, 259 (1978).

14. P. Lagrange, M. ElMakrini, Synth. Metals 2, 191 (1980). 15. B. Warren, X-ray Di&action. 16. 17. IS. 19.

Massachusetts R. Wyckoff. York (1963). R. Vangelisti, R. Vangelisti S. Flandrois, Hauw, Synth.

61

of formation

(1969). Crysfal Structures,

D.

Gerard

and

A.

Addison-Wesley. Vol.

Herold. Reading,

I. Interscience.

unpublished results. and A. Herold. Carbon 14, 333 (1976). J.-M. Masson, G.-C. Rouillon. J. Gaultier Metals 3. 1 (1981).

New

and C.