Materials Letters 30 (1997) 363-368
Preparation of Co and Fe nanocomposites A. Basumallick
a-*, K. Biswas b, S. Mukherjee
a Department ofMetallurgical b Department
Received 8 July 1996; revised 6 September
in SiO, matrices b, G.C. Das b Howrah
University, Calcutta 700032,
Cobalt and iron nanocomposites have been prepared by in situ reduction of metallic salts in SiO, gel matrices. The presence of nanosized metal phases has been established chemically as well as by selected area diffraction and transmission electron microstructural analyses. Studies on temperature and time dependent fractional reduction revealed that the reduction kinetics are fast. Keywords:
Fe, Co; Nanocomposite;
In situ reduction;
1. Introduction Glass-metal nanocomposites exhibit remarkable optical [l-3], electrical [4-61 and magnetic  properties. From the view point of exploiting these new class of materials commercially, extensive research work has been carried out on the preparation of these materials through different routes [8- 121. Recently preparation of glass-metal nanocomposites via the sol-gel route has been fairly well established and has gained considerable importance. Through this route the nanocomposites were prepared by reducing metallic salts in the gel matrix by passing H, from outside . However, this method suffers from some serious limitations. Firstly, the reducing gas should diffuse through the interconnected pores of the gel matrix to cause reduction of the metallic salts. This
00167-577X/97/$17.00 PII SO167-577X(96)00223-6
tends to slow down the kinetics of the process. Secondly, during reduction at high temperatures with H, stringent precautionary measures have to be observed. To overcome the above limitations, we have developed a process to reduce NiCl, in the SiO, gel matrix by in situ generation of H, [lo-121. While preparing nanocomposites through this route one can expect significant advantages. Firstly, it eliminates the possible hazards of reducing metal-chlorides in the gel matrix at high temperatures by passing H, from outside. Secondly, as compared to reduction by passing H, from outside, in the in situ reduction, H, is generated right at the reaction sites, therefore, the reaction rate of the reduction is expected to be faster. The present work deals with the detailed study of reducing CoCI, and FeCl, in the SiO, gel matrix by the in situ generation of H,. This has been accomplished by incorporating the reductant in the gel matrix along with the metal choloride. On heating the gel under a N, atmosphere, the incorporated
0 1997 Elsevier Science B.V. All rights reserved.
et ~1. /Materids
reductant decomposes and generates H? which in turn reduces the metal chloride. The microstructure of the heat-treated gel was studied by a transmission electron microscope (TEM) and the presence of metallic granules was confirmed by the selected area diffraction (SAD) analysis. In addition, investigations have also been made to quantify time-dependent fractional conversion at different temperatures for all the nanocomposites prepared.
CoCl Z and FeCl, which would yield 5 wt% and 10 wt% of Co and Fe in the heat-treated gel matrix were prepared separately. Weight percent was calculated with respect to the amount of SiO, present in the heat-treated gel. These gels were prepared in the following manner. (i) A known amount of metal chloride along with 50% excess of stoichiometric glucose required for the reduction of the metal chloride was dissolved in 5 cm3 of double distilled water. This solution was added to 10 cm3 of ethyl alcohol (C? H,OH) and a homogeneous solution was made. (ii> A homogeneous solution of 10 cm’ C, H,OH and 5 cm3 of tetraethylorthosilicate (TEOS) was also prepared separately. (iii) The first solution was added dropwise to the second solution under continuous stirring with a magnetic stirrer. The resulting solution was left for 16 h gelling at 50°C. Differential thermal analysis (DTA) and thermogravimetric analysis (TGA) of the gels containing CoCI, and FeCl, which would yield 5 wt% of metal on reduction were carried out under pure N, at 1 atm pressure in a thermal analyzer (Shimadzu, DT-40). N, gas was bubbled through an alkaline pyrogal101 solution and then made to pass through a packed bed of anhydrous calcium chloride. The resulting gas free from oxygen and moisture was passed slowly through an impervious tube of an electrical heating furnace. The desired temperature was maintained by a PID controller with an accuracy of f 1°C. The constant heating zone inside the impervious tube was identified by a chromel-alumel thermocouple. The gel sample containing a known amount of metal chloride and glucose was placed in a boat and intro-
Letters 30 (1997) 363-368
duced into the constant heating zone of the tube. All the gel samples were heat treated for 1 h. HCl vapour was generated during the course of the heat treatment. The resulting gas mixture (N2 + HCI vapour) was made to pass through a known volume of double distilled water contained in absorbing towers. The pH values of the resulting HCl solution were recorded as a function of heat-treatment time. The pH value of the solution yield [H+] in the solution from which the fraction of metal chloride reduced was calculated. This process was repeated for temperatures of 900°C and 950°C for CoCl, and 950°C and 1000°C for FeCl, containing gels. The microstructure and SAD patterns of the heat-treated samples were studied by TEM (JEM-200X JEOL).
3. Results and discussion The in situ reduction of CoCl, and FeCl, to the corresponding metal in the gel matrix was chemically established by the generation and detection of HCl as per the mechanism given below [lo]. On heating, glucose decomposes to carbon and water vapour C,H,,O,=6C+6H,O.
The carbon so formed reacts with water vapour available from the gel matrix and decomposition of glucose to generate H, via the water gas reaction: C+H,O=CO+H,. The Hz so generated in situ reduces chloride present in the pores of the gel HZ + metal chloride = metal + HCl
(2) the metal
The presence of HCI in the product gas mixture was identified chemically by the precipitation of AgCl on treating it with AgNO, solution. Fig. 1 shows the representative DTA and TGA plots under N2 at 1 atm for gels containing only CoCl, and FeCl,. The above gels would yield 5 wt% of metal with respect to the SiO, present in the gel matrix on complete reduction when the calculated amount of the reductant is incorporated in the gel matrix and heat treated under N? atmosphere. The features which are observed from Fig. 1 are: (i) The DTA plot exhibits an endothermic trough for both CoCl, and FeCl, containing gels accompa-
Letters 30 (1997) 363-368
oL--__-J ---- 5 wt% co -
Time (min) _-_-co
Fig. 2. Effect of time and temperature CoCl, containing silica gel.
on the in situ reduction
Fig. 1. Thermal analysis (DTA and TGA) in a N? atmosphere of Co and Fe chloride containing gels which would yield 5 wt’% metal on reduction.
nied by weight loss in the corresponding TGA plot. This may be attributed to the removal of water and alcohol present in the pores of the gel matrix. (ii) A broad exothermic peak for gels containing CoCl, and FeCl, was also observed which can be attributed to the carbonization of the alkoxy group followed by oxidation [ 131. Unlike CoCl, containing gels, FeCl, containing gels exhibit a short exothermic peak at 230°C before broadening at 280°C. It is to be noted that although the experiment was conducted under a constant flow rate of N?, the possibility of the presence of O2 in the pores of the gel cannot be eliminated completely. Therefore, the occurrence of the sharp exothermic peak at 230°C for FeCl 3 [ 141 may be attributed to the partial oxidation of FeCl, as follows: $FeCl,
For both CoCl, and FeCl, containing gels, the TGA plot exhibits a continuous weight loss at higher temperatures which accounts for the water loss from polycondensation of the gel matrix. Thus, water vapour continuously liberated from the gel matrix seems to be adequate for the water gas reaction. The effects of time and temperature on the in situ reduction of the CoCI, and FeCl, containing gels
‘.OY 0.9 -
5 wt% Fe
Fig. 3. Effect of time and temperature FeCl 3 containing silica gel.
on the in situ reduction
366 Table I Comparison Co/SiOz
of the observed (Sample
rt ul. /Matericrl.\
(in A) with standard
d hl I
I .2915 0.9546
ASTM data for the Co/SiOz
observed d hi,
2.0268 I, 1708
are shown in Fig. 2 and Fig. 3. From the plots it is observed that the reaction was nearly complete within 15 to 20 min from commencement. This indicates that the rate of the reaction was very fast and in agreement with our expectation. From the plot it can also be observed that the fractional conversion remains almost identical during the initial stages irrespective of temperature. Maximum conversion occurs within the first five minutes from commencement of the reaction. This is believed to be due to an enormous increase in the number of reaction sites and high energy input due to the reductions occurring at high temperatures. However, full conversion was not observed for both CoCl? and FeCl, containing gels. This is attributed to the fact that during the progress of the reaction at high temperatures, water vapour available for carrying out the water gas reaction is removed at a rapid rate. This is due to the decrease in volume of HZ generated and secondly, to the rapid removal of unreacted HZ from the reaction sites. Table 1 shows the typical sets of data pertaining to the interplanar spacings cd,,,) as calculated from
Table 2 Heat-treatment
schedule and particle diameters
the selected area diffraction rings for the Co/SiOZ and Fe/SiO, nanocomposites. The standard values (ASTM) for the corresponding metallic species were also shown for comparison. The observed interplanar spacings match reasonably well with those of the standard values [ 161. It is to be noted that some of the d,,,, values listed in Table 1 arise due to the presence of respective oxides, namely Fe,O, and Co30,. However, the amount of the oxides seems to be rather small because the intensities of these rings were weak. Therefore, from the SAD pattern analysis, the presence of metallic Co and Fe in the SiO, matrix was confirmed. Table 2 summarises the mean particle diameter (d) and the standard deviation (5,) of particle sizes for the different heat-treated samples. The average particle diameters for Co and Fe are respectively in the range 92 to 110 nm and 14 to 60 nm. It is to be noted that both, for Co and Fe, the standard deviations are of the order of the average particle sizes. This means that there is a wide distribution in particle sizes in the nanocomposites. Fig. 4a shows the electron micrograph of sample 1A in the Co/SiO,
of different composites
after heat treatment.
S, is the standard deviation of the particle sizes
Co/SiO, Co/SiO, Co/SiO, Co/SiO, Fe/SiO; Fe/SiO, Fe/SiO, Fe/SiOi
IA IB 2A 2B 3A 3B 4A 4B
5 5 IO IO 5 5 10 IO
900°C (60 min) 950°C (60 min) 900°C (60 min) 950°C (60 min) 950°C (60 min) 1000°C (60 min) 950°C (60 min) 1000°C (60 min)
100.08 110.92 88.96 91.60 61.10 45.00 5 1.60 13.70
S, (nm) 74.41 67.28 65.40 42.00 46.20 22.80 23.00 8.70
Letters 30 11997) 363-368
system. The corresponding SAD pattern is shown in Fig. 4b. Spotted rings in the SAD pattern were observed. This can be attributed to the presence of large particles which may arise by grain coarsening (growth) due to prolonged holding at high reduction temperatures. Fig. 5a shows the electron micrograph of sample 4B in the Fe/SiO, system. The corresponding SAD pattern is shown in Fig. 5b. Clear non-spotty diffraction rings in the SAD pattern are observed. In sharp contrast to the Co/SiO, composites, the mean particle diameter is observed to decrease with increasing temperature, Table 2. This effect can be attributed to the low boiling point (280°C) of FeCl, . The simultaneous loss of unreacted FeCl 3 by evaporation from the SiO, matrix reduces the chance of grain coarsening. The higher the temperature of reduction
Fig. 5. (a) Electron micrograph of sample composite. (b) SAD pattern of (a).
4B in the Fe/SiO,
the higher the loss of FeCl, and so the lower will be the chance of grain coarsening. Therefore, unlike COCl,, the average diameter of metallic iron decreases with higher temperature of reduction. Thus, the presence of nanosized metallic phase in the SiO, gel matrix was established chemically by the detection and identification of HCl and has been further substantiated by SAD pattern analysis.
Fig. 4. (a) Electron micrograph of sample composite. (b) SAD pattern of (a).
IA in the Co/SiO,
On the basis of the results obtained during the present investigation the following conclusion have been drawn:
et al. /Materials
(i) The in situ reduction of metal chlorides, e.g. CoCl, and FeCl,, in the SiO, gel matrix is feasible and has been established by chemical as well as by SAD pattern analysis. (ii) The average particle diameter for both Co and Fe in the nanocomposites is of the order of 10 to 100 nm. (iii) In the Co/SiO, composite, grain coarsening has occurred which has been attributed to the prolonged holding of the sample at the high reduction temperature. (iv) In contrast to the Co/SiO, composite, the Fe/SiO, composite exhibits decreasing Fe particle diameters with increasing reduction temperature. This is due to simultaneous loss by evaporation of unreacted FeCl, during reduction. (v) The reduction progresses at a rapid rate due to the enormous increase in the number of reaction sites as provided by the simultaneous in situ generation of hydrogen through the matrix. For both Co/SiO, and Fe/SiO, composites, maximum conversion occurs within the first 5 min, after the commencement of reaction.
Letters 30 (I 9971 363-368
References [I] D. Chakravorty,         [IO] [Ill  [I31 
The authors thank the Council of Scientific Industrial Research, New Delhi for sponsoring research on nanocomposites.
A. Shuttleworth and P.H. Gaskell, J. Mater. sot. 10 (1975) 799. C.G. Granqvist and 0. Hunderi, Phys. Rev. B 16 (1977) 3513. G.C. Das, R. Das and D. Chakravorty, Bull. Mater. Sot. 5 (1983) 277. A. Chatterjee and D. Chakravorty, J. Phys. D: Appl. Phys 22 (1989) 1386. G.C. Das, T.K. Reddy and D. Chakravorty, J. Mater. Sot. 13 (1978) 3211. A. Chatterjee and D. Chakravorty, J. Phys. D: Appl. Phys 23 (1990) 1097. S. Dutta, D. Bahadur and D. Chakravorty. J. Phys. D: Appl. Phys 17 (1984) 163. C.G. Granqvist and R.A. Buhrman, J. Phys D: Appl. Phys. 47 (1976) 2200. S. Dutta, S.S. Mitra, D. Chakravorty, S. Ram and D. Bahadur, J. Mater. Sot. Lett. 5 (1986) 89. G.C. Das, A. Basumallick and S. Mukherjee, Bull. Mater. Sot. 13 (1990) 255. G.C. Das, A. Basumallick, K. Biswas and S. Mukherjee, Bull. Mater. Sot. 16 (1993) 317. A. Basumallick, G.C. Das, K. Biswas and S. Mukherjee, J. Mater. Res. 10 (1995) 2938. G.J. Brinker, K.D. Keefer. D.W. Schaefer and CS Ashley, J. Non-Cryst. Solids 48 (1992) 47. A. Volsky and E. Sergievskaya, in: Theory of metallurgical processes. 2nd Ed. (MIR Publishers, Moscow, 1978) p. 344. R.H. Perry and C.H. Chilton, in: Chemical engineer’s handbook (Kogakusha Ltd, Tokyo) pp. 3- 12. Selected powder diffraction data for metals and alloys databook, Vols. I and II, 1st Ed. (International Center for Diffraction Data, Pennysylvania, 1978).