Influence of the Composition of the Boroning Mixture on the Dimension Change of Pressed and Boroned Samples from Iron Powder

Volume changes occur during sintering and chemical-thermal treatments of metal powder samples. The results of the investigation of the volume change of pressed and boroned samples from an iron powder, depending on the mixture composition used for the boroning process, are presented in this paper. The basic mixture, used for boroning of the investigated samples from iron powder, is modified by the addition of activators with different chemical compositions and in different concentrations, of up to 4 wt %. Mixtures with ammonium bifluoride, ammonium chloride and boron potassium fluoride were investigated. The research results and the mathematical modelling enable the choice of mixture compositions for boroning based on the volume change given in advance.


Introduction
The chemical-thermal treatment of products obtained by powder metallurgy processes is quite a new method and only few papers deal with boroning of iron powder compacts [1-3, 11, 12].According to these researches, it is necessary to predict the possibilities of obtaining boroning layers within the iron powder compacts, as well as the possibilities of sintering in the presence of chosen activators.
A contribution to the research, in relation to the dimension change by boroning of iron powder compacts through the variation of composition mixtures for boroning, (enabling diffusing layers having different depth), is described in this paper.This approach should help sort out the changes which arise due to boroning [1][2][3], as well as from the changes which arise due to activated sintering [4][5][6][7][8].The mass transport, during sintering of metal powders, leads to the creation of greater contacts between particles of the powder, and finally, to the shrinkage and the increase of the relative density.The transfer of matter, which leads to shrinkage, is caused by the diffusion or viscous flow in the solid state.These processes are usually directionally dependent and therefore the value of the linear shrinkage is different along the different axes of the powder compacts.Investigation of the linear shrinkage gives some information about the major direction of matter flow during sintering, and accordingly, offers better understanding of the mechanisms which lead to shrinkage [4,5].
The mechanism and kinetics of sintered materials of diffusing layers differ considerably from those of cast materials [1,[9][10][11][12].The kinetics of the diffusing layers formation, the structure, properties and the phase composition depend, to a large extent, on the structural characteristics of the sintered materials.The porosity of sintered materials has a significant influence on the diffusing layers' quality.
As for the chemical-thermal treatment of materials with closed porosity, the intrusion of active atoms from saturating environment into the inner part of the sample, is excluded, so the diffusing layer formation on sintered materials differs from the one relating to the casting materials, because of the presence of closed porosity, crystal structure defects, the crystal lattice deformity, larger length of grains and subgrain boundaries as well as because of the presence of admixtures.Due to the mentioned facts, the diffusion process is accelerated in sintered materials.
The process of saturation, as is the case with the greater number of elements, causes lessening of sections and closing of pore channels due to diffusion layer formation on their surface and additional sintering as well.Therefore, it is possible to conclude that sintered materials are transited very quickly from the semi contiguous porosity to closed porosity materials.
When considering open porosity materials, the active environment penetration through open pores into the inner part of the compact has a strong influence on the speed rate of diffusing layer formation and their structure.When examining the saturation of open porosity materials, one should keep in mind that the initial porosity changes in the saturation process.The diffusing layer formation near the open pores, occurs along with the volume enhancement; therefore, the surface section of the open pores channel is diminished during the chemical-thermal treatment.In fact, this hinders the penetration of the saturating element into the inner part of the compact.Articles published on these problems, allow several assumptions [2,3,13]: 1.The closing of pore channels supports the saturation with elements naturally, similar to Fe (Cr, V, Cu, Ni, Mn).
2. Saturation with elements by a natural process, is considerably different from saturation with Fe (Al, Si, S) and does not make the complete closing of pore channels possible but only lessens the section thereof.
3. The closing of pore channels is intensified by a fluid phase formation during the saturation process as well as by the formation of chemical compounds.During the process of chemical-thermal treatment, due to the proximity of chemical-thermal treatment to the sintering regime, the process of compacts' sintering occurs.The atmosphere formed during the process of saturation considerably stimulates this process, due to which, with the application of gas saturation methods, a "non-porous" transitional zone can be formed under the diffusing layer, which has a significant influence on the final properties of the sample.The research, presented in this article, showed that the boroning process is possible for nonsintered samples.

Experiment
Application of the simplex experimental plan with fifteen experimental points, schematically shown in Fig. 1, presented in the article, is based on choosing the optimal conditions of boroning for the pressed iron powder samples.
The samples obtained from iron powder NC 100.24 (Höganäs) by pressing under 400 MPa, were subjected to boroning in a solid mixture with boron carbide.The samples had rectangular forms, with the cross-section dimensions of 31 mm × 12 mm.Our own experience was used in selection of the boroning mixture, so the best results of the boride layer depth and quality were achieved using a mixture with boron carbide (B 4 C) and activators [6,7].NH 4 HF 2 , NH 4 Cl, and KBF 4 were used as activators in the boroning process.The content of the basic components in the compound prepared for boroning was constant, but the activator's content was variable.The percentage share of the activator was different and it was in the range of 0 to 4 wt %.As for the mathematical model, a polynomial of the fourth degree was used, given by relationship (1).The regime of boroning was the same for all samples: the temperature was 950 °C and the process time was 4 hours.The plan of the experiment as well as the experimental results of the volume change and the layer's depth are given in Tab.I.The linear dimensions of samples, before and after the process of boroning on the basis of which the volume change was determined, were measured with the precison of ± 0,01 mm.The boride layer depth was assessed by a device for hardness measurement, having precision of ± 0,5 µm.For each sample the cross-section was measured five times and the average values were calculated.The obtained results given in Tab.I represent the average values of the repeated readings.

Examination Results
The investigations were carried out using a simplex plan of the fourth degree with fifteen experimental points.It has been used for choosing the optimal boroning conditions for the iron powder compacts.The mathematical model is given by a polynomial of the fourth degree, since incomplete cubic and cubic models are inadequate.The results of the explained simplex plan are presented in Tab.I.
Explanation of the experimental data by the simplex method [14, 15] reveals that there is a strong dependence of the volume change on the variation of the composition mixtures for boroning (1).Fig. 2 brings out this influence more clearly.The volume change of boroned samples follows from equations:  • The volume change of samples is given with the equations, where V ⋅ B represents the volume of samples after boroning, V 0 -the volume of iron powder compacts.
The isovolume lines of boroned samples are drawn and presented in the coordinate system of concentration -composition of activators (Fig. 2).After the dispersion analysis, which gave satisfactory results, the numerical data were checked by the mathematical model at control points K 1 (x 1 = 0.71; x 2 = 0.11; x 3 = 0.17) and K 2 (x 1 = 0.08; x 2 = 0.82; x 3 = 0.1), and it was confirmed with a probability of 99 % that the mathematical model was adequate.
Depending on the amount of activators, the volume change was in the range of -1.581 to 4.449 %.During the process of chemical-thermal treatment, due to the proximity of the chemical-thermal treatment to the sintering regime, the process of sintering compacts occurs.After boroning, the dimensions can remain the same or can change due to shrinkage or growth of samples.It can be seen from Tab.I that the influence of boroning the mixture composition on the volume change of samples is dual, e.g. in some samples shrinkage occurred, while in other samples the increase of dimensions (e.g.growth) occurred.For samples where shrinkage occurred, the compactness of the boride layer increased, Fig. 3 and Tab.I. Fig. 3 The microstructure of a boride sample obtained from Fe powder by pressing under 400 MPa.The sample was boroned by the basic mixture with 4 % KBF 4 (×500).
It is essential to determine the dependence of the obtained boride layer depth from changes of sample dimensions, as well as the relationship between mass changes (Δm) and volume change (ΔV).Fig. 4 and Fig. 5 represent these dependences.Boride formations from the very start of the boroning process were not attended by adequate changes of the crosssection dimension (Δh).The dimension increase was noticed from the moment when the layer of about 132 μm was formed.The increase of mass during boroning was noticed from the start of the process, while the volume change was registered from the moment the mass was grown to about 0.58 g (Fig. 5).
Within the framework of this article, a metallographic analysis of the quality of boride samples was conducted, the boride layer depth was measured and the porosity was determined with vacuum methods.On the basis of the obtained results, the following can be concluded.The samples obtained by pressing iron powder, approximately with the same density, after the process of boroning, don't have the same porosity.The porosity differs by the pore density, shape and dimensions (Fig. 6).These differences can be explained by the mere dependence of the composite mixture for boroning on sintering during the process of boroning.The added activators are being dissolved at the boroning temperature and their products activate the process of sintering.During the process of activated sintering, the spheroidization of pores is being rated and the mass transport during sintering of metal powder leads to the formation of closer contacts between powder particles; so the porosity occurs in the form of particular pores distributed on the compact surfaces (Fig. 7a).On the samples, boride layers of different depths have been obtained, composed of boride Fe 2 B of different quality (porosity, the bond with basic material), (Fig. 2 and 7).The layers' depth changed with the various quantities of added activators.
Therefore, on the sample boroned by the basic mixture having 4 % of NH 4 HF 2 and 4 % of KBF 4 , it is possible to see crystals of boride Fe 2 B, which penetrate, in a wedge shape into the basic material (Fig. 7b).The porosity and oxides prevent the boride growth.The depth of the boride layer amounts to 118.0 µm.
The boroned sample by the mixture of 4 % of NH

Fig. 1
Fig. 1 Schematic presentation of the simplex experimental plan 2

Fig. 2
Fig. 2 Dependence of volume change of pressed and boroned samples on the boroning mixture composition (o -experimental points; x -control points)

δFig. 4
Fig. 4 Dependence of the boride layer depth (δ) on the sample dimension change (Δh) (The coefficient of correlation equals to 0.46)

Fig. 5
Fig. 5 Dependence of the mass change (Δm) on the volume change (ΔV) (The coefficient of correlation equals to