A Comparative Study of Spark Plasma and Conventional Sintering of Undoped SnO2 Sputtering Targets

SnO 2 ceramics were fabricated by spark plasma sintering (SPS) and conventional (pressureless) sintering techniques by using undoped submicron SnO 2 powders. The effect of sintering temperature and dwell time on the densication behavior, phase evolution and microstructural development of sintered ceramics were investigated. The relative density of SPSed ceramics increased when dwell time was raised from 1 to 10 min at 950ºC. However, full densication was prevented by the decomposition of SnO 2 to Sn and O 2 (g). The decomposition starts after ~ 10 min at 950ºC. In parallel to this observations, as sintering temperature increases, amount of the elemental Sn in agglomerated form increases. On the other hand, the relative densities of conventionally sintered ceramics (at 1200ºC-1400ºC) were relatively low (i.e., 63 % relative density), and abnormal grain growth was observed due to the shift in sintering mechanisms to evaporation-condensation as a dominant mechanism. Since the undoped SnO 2 ceramics, SPSed at 950°C for 5 min under 30 MPa exhibit 93 % relative density, high chemical purity, homogeneous grain size distribution and smaller average grain size, they demonstrate great potential as sputtering targets for production of high-quality thin lm gas sensors.


Introduction
Tin oxide (SnO 2 ) is a critical electronic material which has been widely used in an extensive range of applications such as gas sensor [1], varistor in electronic devices [2], electro/photocatalyst [3,4], transparent conductive oxide lms [5] and optoelectronic devices [6]. The most important application eld of SnO 2 material is in metal oxide semiconductor (MOS) based gas sensors. The MOS sensors are the most preferred and currently used devices among the other sensing mechanisms such as catalytic, electrochemical, and NDIR (Non-Dispersive Infrared) types to detect the pollutant gases (CO, CO 2 , NO x , SO x , VOC s, etc.) results in diminishing the air quality. SnO 2 based materials can be deposited into the gas sensing devices in the form of thick or thin lms [7].
Thick lms are composed of porous SnO 2 in a thickness of 1 to 10 microns produced from the slurry form. Thin lms, on the other hand, have a thickness of 10 nm to 1 micron. Several methods were used to prepare SnO 2 based gas sensors, such as thermal evaporation [8], dip and spin coating [9], and hydro/solvothermal method [10,11]. For industrial aspects, sputtering is a widely accepted route for preparing SnO 2 thin lms used in gas sensing applications among these methods. Sputtering process can be performed from either metallic or ceramic targets. Ceramic targets are preferred to produce thin lms with higher quality and reproductively. They have homogeneous optical and electrical properties along with the material, whereas the metallic one is cheaper and more accessible in the market and has higher thermal conductivity. One of the well-known problems during sputtering is the formation of nodules on the target material [12,13]. The nodules formed by the arching during the sputtering can result in the uncontrolled highly defective lm. The arching is mainly related to the non-homogeneity in the chemical, physical, microstructural, and electrical properties of the sputter target. In addition to that, inhomogeneous grain size, a minimum amount of pores and higher density prevent the formation of nodules. The formation of particles during the sputtering process is one of the signi cant problems which originate from nodules on the target surface. To overcome the nodule and related causes during the sputtering process, target material should be manufactured with higher density (> 92 % of SnO 2 relative density ), higher chemical purity (> 99.95 wt.% of SnO 2 ), homogeneous distribution of cations and oxygen mainly in the surface and higher thermal and mechanical stability to reduce the chance of target cracking. Even ceramics with a density of greater than 98% of its relative density and high electrical conductivity are strongly required for DC magnetron sputtering systems employed in the industry [14].
It is crucial to produce dense SnO 2 sputtering targets with homogenous grain size distribution and high chemical purity which provides low electrical resistivity and high thermal and mechanical stability due to getting relatively higher performance in the sputtering process. However, it is di cult to achieve full densi cation by pressureless-assisted sintering due to the decomposition of SnO 2 (s) to SnO(g) above 1200ºC where evaporation-condensation takes place [15]. Therefore, it is concluded with only coarsening of tin oxide ceramics with very low densi cation (60 % relative density) [14]. Leite et al. [16] investigated pressureless-assisted sintering kinetics of ultra ne undoped SnO 2 powder under atmospheric conditions.
The results showed that the decomposition rate of SnO 2 increases above 1300ºC with the decreasing particle size of starting SnO 2 powder (from approximately 100 to 25 nm).
Possible technique to produce tin oxide-based ceramics with a higher density is pressure-assisting sintering [14]. Spark plasma sintering (SPS, FAST) is one of the most favorable sintering techniques in which dense polycrystalline materials are produced in bulk form through powder metallurgy under applied pressure. In comparison with the conventional methods and other sintering techniques such as hot press (HP) and hot isostatic pressing (HIP), SPS is fundamentally different in terms of working principle, which is based on the generation of high pulsed electric current. Thus, both electrical eld and joule heating are provided along with uniaxial pressure. This phenomenon comes with some advantages such as; high-speed diffusion, full densi cation, elimination of impurities on the surface of particles, reduction of sintering temperature, shorting of sintering time, and improving mechanical and electrical properties [17][18][19].
A few types of research in the literature focus on the pressure-assisted sintering of undoped SnO 2 powders. Yoshinaka et al. [20] [21] focused on the spark plasma sintering of undoped SnO 2 powders using commercial undoped SnO 2 powder (60 nm crystallite size), and they investigated the variation of the relative density and the weight loss with the sintering temperature. It is reported that weight loss began to occur above 1050ºC. The sintered ceramic by spark plasma sintering at 1050ºC for 5 min under 37.5 MPa exhibited the maximum relative density (around 95 %). That is the only study in the literature based on the spark plasma sintering of undoped SnO 2 powders. Unfortunately, the decomposition of SnO 2 did not discuss in detail; there was no data regarding the phase evolution and microstructural development. Park et al. only focused on the microstructure analysis of SnO 2 ceramic sintered at 1050ºC in terms of the formation of pores and the homogeneity of microstructure. Delorme et al. [22] stated that undoped SnO 2 ceramics were produced by SPS in order to improve transport properties that are sensitive to doping elements of oxides (CoO, MnO 2 , CuO, ZnO, etc.). In this study, they achieved almost 95 % relative density at 950ºC for 10 min under 100 MPa. Delorme et al. mainly focused on the electrical and thermal conductivity of SnO 2 ceramics, which were obtained at different SPS conditions. However, this study does not re ect detailed microstructural development among thermodynamical approaches to understand densi cation behaviour of SnO 2 ceramics under SPS conditions. A correlative assessment of chemical purity, density, homogenously distributed grains, and small grain size of sputtering targets is crucial to both investigate and improve the performance of sputtering targets. To eliminate this shortage in the literature, in the present study, it is aimed to develop a fundamental understanding of de ning ideal microstructure and its effect on nal target properties based on densi cation behavior, phase evolution, and microstructural development in order to get higher performance and e ciency during the sputtering process and achieve higher quality of as-produced thin lms as well. Therefore, in this paper, the effects of process conditions on densi cation behavior, phase evolution and microstructural development of SnO 2 ceramic sputter targets obtained from undoped submicron commercially available SnO 2 powders were systematically investigated as a function of sintering temperature (850-1050°C) and time (1-10 min), while the particle size of starting powder, external pressure, atmospheric conditions, and heating & cooling rates keep constant. Furthermore, the results obtained from SPSed ceramics were compared to conventionally sintered SnO 2 ceramics and commercial SnO 2 targets. to the measurement. The average particle size (i.e., equivalent spherical diameter), assuming that the particle is spherical and nonporous, was calculated from the BET surface area via equation (SS = 6/ρ.D where SS is speci c surface area of the particles in m 2 /g, ρ is theoretical density of the particle and D is the particle size).
The qualitative phase analysis was performed using an x-ray diffractometer (XRD; MiniFlex600, Rigaku).
Particle size and distribution were measured using a scanning electron microscope (SEM; Supra 50VP, Zeiss).
Chemical purity of the SnO 2 was determined by x-ray uorescence (XRF, Rigaku ZSX Primus).

Sintering of undoped SnO 2 ceramics
Densi cation of SnO 2 submicron powder without any sintering additive was achieved using spark plasma sintering furnace (HPD-50, FCT GmbH, Germany) under a vacuum atmosphere. The process parameters such as sintering temperature and holding time were optimized to fabricate dense SnO 2 ceramics. During all sintering cycles, heating rate (200ºC/min), cooling rate (switching power off, approximately 600ºC/min), and external pressure (30 MPa) were kept constant. 4 grams of powder were loaded into a graphite die with an inner diameter of 20 mm. Graphite foil was incorporated on the inner surface of graphite die to prevent the reaction between the graphite die and powder. The temperature was increased by the controlled electrical current with the pulse timing of 12:2, and it was monitored by optical pyrometer inside the graphite punches. In order to compare spark plasma sintering with conventional sintering technique, undoped SnO 2 ceramics were also sintered by presureless method (Carbolite Lab Furnace) under conditions (1200ºC, 2 h, 1 atm, 5°C/min) optimized in a prior study [23]. The samples with 50 mm diameter were prepared via uniaxial pressing under 200 MPa for conventional sintering.
Using the submicron starting powder lieS in its high speci c surface area and high surface curvature, which improve the sintering rate. Henrring equation explains well the importance of particle size on densi cation. For instance, when the particle size is 2 times higher, the sintering time is 16 times, where the grain boundary diffusion is dominant [24,25]. Leite et al. [16] investigated the grain growth kinetics of nanometric undoped SnO 2 powder in the temperature range of 500-1300ºC. Experimental results showed that the mass transport mechanism was controlled by surface diffusion between 500-1000ºC. The thermodynamic driving force for sintering is the reduction in total interfacial energy by diffusion. The differences in bulk pressure, vapor pressure, and vacancy concentration depending on interface curvature yield material transport in terms of kinetics [24]. Different mass transport mechanisms occur during sintering, such as lattice diffusion, grain boundary diffusion, surface diffusion, viscous ow, and gasphase transport. Some of these sintering mechanisms contribute to densi cation, whereas others (e.g. surface diffusion) lead to grain growth without densi cation, which are called non-densifying mechanisms. The dominant densi cation mechanism depends on experimental and compact conditions such as particle size, neck radius, temperature, time, and pressure [24,26]. In order to limit the surface diffusion as a non-densifying mechanism, a high heating rate of 200°C/m was constantly applied during spark plasma sintering of SnO 2 ceramics.

Characterization of the sintered ceramics
After removing the graphite foil layer on the surface of SPSed samples (designated as SPS-SnO 2 ), the bulk densities of sintered ceramics were determined by the Archimedes method. In order to calculate the relative density, the density value obtained by Archimedes' principle was divided to 6.95 g/cm 3 (SnO 2 theoretical density). SnO 2 ceramics were crushed and ground under 63 µm. Qualitative phase analysis was performed using XRD (MiniFlex600, Rigaku) with a scan speed of 2°/min at 40 V and 15A. Phase determination was carried out by the Rietveld method using Material Analysis Using Diffraction (MAUD) software. The fractured surface (cross-sectional area) of samples was polished by the automatic polishing machine (TegraPol-25, Struers). Then, all polished samples were thermally etched at a temperature of 200ºC lower than sintering temperature to observe ne microstructure. The microstructure analysis and quantitative elemental analysis carried out using SEM (Supra 50VP, Zeiss). In addition, undoped SnO 2 ceramics fabricated by conventional technique (designated as CS-SnO 2 ) and a commercial SnO 2 target purchased from a target manufacturer (designated as CM-SnO 2 ) were also characterized by the same techniques.
3 Results And Discussion 3.1 Characterization of the starting SnO 2 powder XRD pattern and SEM image of starting powder are illustrated in Fig. 1 and Fig. 2. The powder is composed of only Cassiterite (SnO 2 ; JCPDS #41-1445), a Rutile type of crystal structure. According to the SEM image of starting powder, the particle size of SnO 2 particles is ranging from 100 to 500 nm with irregular morphologies. The average particle size measured from the SE images is 150 nm. The speci c surface area and calculated equivalent spherical diameter of starting powder particle size by the BET method are 7.6 m 2 /g and 110 nm, respectively. The chemical purity of the starting SnO 2 powder measured by XRF is shown in Table 1. The SnO 2 powders have high purity and are suitable for the sputtering process without any puri cation before sintering. The loss on ignition value is neglected due to its shallow value (< 0.1 %).  Sintering parameters and properties of sintered ceramics are given in Table 2. For constantly applied pressure, heating rate, sintering temperature (950ºC), with the increment of dwell time from 1 to 10 min, relative density of SPSed SnO 2 increased from 82 to 94 %. The relative density values of SPS1-SnO 2 and SPS2-SnO 2 are found below 92 %, which signs the limited densi cation into the second step. The relative density gradually increases from 82 to 93 % with the increment of dwell time from 1 min to 5 min. When the dwell time increases from 5 min to 10 min, the relative density increases only 2 % (up to 94 %), and average grain size reduces unexpectedly. It is consistent with theoretical knowledge that the required sintering time for densi cation exponentially increases while the relative density reaches maximum point [24]. After increasing sintering temperature from 950ºC to 1050ºC with constant dwell time (10 min), relative density remains same (94 %), and average grain size increases. The relative density of conventionally sintered SnO 2 is same (63 %), although sintering temperature increases from 1200ºC to 1400ºC under constant pressure, heating rate and dwell time. Commercial target has the same relative density as conventionally sintered SnO 2 .
XRD patterns of sintered ceramics are given in Fig. 3. According to the obtained results, all SPSed ceramics is composed of SnO 2 (Cassiterite; JCPDS #41-1445) as a monolithic crystalline phase except the SPS5-SnO 2 sintered at 1050ºC that contains SnO 2 and elemental Tin (Sn; JCPDS #04-0673). Thus, it con rms that a reasonable amount of elemental Sn is formed at 1050ºC and solidi es during cooling.

Effect of sintering temperature on microstructure development
As stated before, the relative density remains the same even though the sintering temperature is raised from 950ºC to 1050ºC, and only SPS5-SnO 2 contains elemental Tin (Sn; JCPDS # 04-0673) as a secondary phase. Rietveld analysis pattern obtained from powder diffraction data of SPS5-SnO 2 is given in Fig. 4. According to phase determination carried out by the Rietveld method, elemental Tin content in SPS5-SnO 2 is ~ 3 wt. %. There are two possible chemical reactions for the decomposition of SnO 2 which are given in (1) and (2). ΔG-Temperature diagrams for the decomposition reactions, obtained using FactSage thermochemical software ( Hence, the decomposition reaction was thermodynamically favorable even at 1050ºC in comparison to 1650ºC, which is expected to occur without any electric eld effect. As a result, solid SnO 2 was able to decompose into Sn (l) and O 2 (g). Following to thermal process, the solidi ed Sn in the structure could be detected by XRD as a minor crystalline phase.
SEM images (BSD mode) of SPS4-SnO 2 and SPS5-SnO 2 are given in Fig. 6. It is observed that the grain size of SPS4-SnO 2 is small (about 200-300 nm), grain morphology is nearly equiaxed, and grain size distribution is homogenous. Also, elemental Sn in trace amount is formed as a decomposition product, which could not be clearly observed by XRD. As seen in Fig. 6(a), Sn particles can be seen on the surface of SnO 2 , having bright contrast, since decomposition of SnO 2 particles starts on grain boundaries. When the sintering temperature of SnO 2 is increased further up to 1050ºC, the grain growth is accelerated by means of agglomerated Sn grains distributed non-homogenously. In addition to that, higher content of decomposed elemental Sn can now be easily detected by XRD analysis, preventing achieving full densi cation due to condensation-evaporation mechanism. Delorme et al. [23] showed similar decomposition behavior of undoped SnO 2 at 1050ºC for 5 min under 37.5 MPa during SPS.
The temperature and displacement (relative piston travel) curves obtained through spark plasma sintering of SPS4-SnO 2 and SPS5-SnO 2 are demonstrated in Fig. 8. All temperature curves show one stepheating schedule. The displacement curves obtained by the piston motion demonstrate the compaction of the samples during the heating cycle as indicated by the z-axis increment. The displacement values obtained by schedule data, from the beginning of the heating stage at room temperature untill the end of the heating cycle, are 1.79 mm and 2.02 mm for SPS4-SnO 2, and SPS5-SnO 2 , respectively. For these specimens, compaction occurs at an early stage due to applied pressure, expansion is observed above 400ºC because of the increase of temperature, and a high shrinkage takes place above 800ºC until the end of dwell time given at top temperature. It means that the densi cation starts at ∼800ºC and makes a huge contribution to shrinkage (displacement). During the sintering of SPS4-SnO 2 , the displacement rate (ratio of displacement to temperature) slows down at 950ºC after 3 min due to densi cation. However, the rising temperature or the increasing displacement does not always refer to a higher densi cation rate. Even though SPS5-SnO 2 is sintered at a higher temperature (1050ºC) and displacement curves indicate maximum displacement value at 2.02 mm, the relative density values of SPS4-SnO 2 and SPS5-SnO 2 are equal. In this situation, the increase of temperature leads to the formation of elemental Sn in a high amount. A rapid displacement at 1050ºC points out the ejection of Sn from the graphite mold.
Considering the higher theoretical density of Sn (∼7.265 g/cm 3 ) than that of SnO 2 (∼6.9 g/cm 3 ) and a relatively higher content of Sn in SPS5-SnO 2 , the same relative density of SPS4-SnO 2 and SPS5-SnO 2 indicates less densi cation of SPS5-SnO 2 in which Sn exists as a secondary phase.
As known, diffusion is the most crucial sintering mechanism. According to Arrhenius' expression, the diffusion is temperature-dependent. Hence, the increasing sintering temperature increases the sintering rate by promoting atomic diffusion (24,27). For this reason, the effect of sintering temperature on microstructural development of SPSed ceramics has been investigated in this section. However, relative density is limited to 94% even if sintering temperature is raised from 950ºC to 1050ºC. Full densi cation could not be provided owing to elemental Sn as a secondary phase formed through the decomposition of SnO 2 , which is consistent with the literature [21].

Effect of dwell time on microstructure development
The obtained results from previous SPS experiments show that 950ºC is the optimal sintering temperature because higher sintering temperature leads to the formation of secondary phase. Still, SPS4-SnO 2 contains a trace amount of elemental Sn in its microstructure. To investigate the effect of dwell time on densi cation, phase evolution, and microstructural development, the different dwell times (1, 3, 5, 10 min) were applied at an optimal sintering temperature (950ºC) during spark plasma sintering. As stated before, the relative density of ceramics SPSed at 950ºC increases from 82 to 94 % with the increment of dwell time from 1 min to 10 min. According to XRD patterns, they consist of only SnO 2 phase.
It is seen that the grain size of SPS1-SnO 2 and SPS2-SnO 2 is unexpectedly larger than that of both SPS3-SnO 2 and SPS4-SnO 2 . The agglomerated grains in the microstructure of SPS1-SnO 2 and SPS2-SnO 2 are illustrated with red plus signs (+) in Fig. 9. Non-homogenous grain size distribution is associated with the presence of agglomerates with a size above 500 nm. However, there is no signi cant grain growth for SPS3-SnO 2 . A further increase in dwell time up to 10 min develops a microstructure containing grains with smaller average size. Although a holding time of more than 10 min is needed to eliminate isolated pores (complete densi cation), the maximum dwell time is selected as 10 min to prevent further decomposition of SnO 2 and limit liquid Sn content. The grain size reduction above 5 min is associated with the starting of decomposition of SnO 2 into Sn on grain boundaries, where grain growth of SnO 2 is suppressed without affecting the densi cation mechanism. A slight decrease in grain size (~ 50 nm) might be attributed to decreasing volume of reduced SnO 2 into Sn at the beginning of decomposition phenomena. In addition, the particle size of starting powder is 100-200 nm, and the average grain size of SPS4-SnO 2 is found about 200-300 nm. It shows that grain growth is minimal due to different sintering mechanisms of SPS where the pulse current ows through powder particles and then the heating power is dissipating at the contact points of particles [18].
The temperature and displacement curves of SPS1-SnO 2 , SPS3-SnO 2 and SPS4-SnO 2 are demonstrated in Figure 10. The process data of SPS2-SnO 2 could not be collected because of a technical problem. The shrinkage rate remarkably increases above 800°C during the one step-heating cycles. The total displacement value is found 1.54 mm, 1.71 mm, and 1.79 mm for SPS1-SnO 2 , SPS3-SnO 2 and SPS4-SnO 2 , respectively. The displacement (density) non-proportionally increases with the gradual increase in dwell time. As seen in Fig. 10(b) and Fig. 10(c), the full compaction cannot be provided in 5 min because displacement ratio is high, while the displacement reaches a plateau after 5 min where the shrinkage rate is too low. On the other hand, a plateau on displacement curve refers to higher densi cation for SPS4-SnO 2 , whereas that contains elemental Sn in trace amount. In order to produce monolithic SnO 2 ceramics with high chemical purity, 5 min is selected as the optimal sintering time for SPS.

Figure11
The schematic illustration of microstructural development of SPSed SnO 2 ceramics as a function of time and temperature The schematic illustration of microstructural development of SPSed SnO 2 ceramics as a function of time and temperature is given in Fig. 11. To summarize the given knowledge in Sect. 3.2., both increment of dwell time from 5 to 10 min at 950°C and the increasing of sintering temperature from 950°C to 1050°C does not contribute to densi cation of SnO 2 (relative density remains almost the same) because of the evaporation of SnO 2 into elemental Sn followed by condensation of Sn onto SnO 2 grain boundaries that is thermodynamically favorable at 950°C under SPS conditions. The formation of Sn has been rstly seen on grain boundaries on the 10th min of sintering at 950°C. Although dwell time is increased from 5 to 10 min, a slight decrease in average grain size is observed unexpectedly, which might be attributed to decreasing volume of reduced SnO 2 into Sn at the rst moment of decomposition phenomena took place on particle surface under external pressure (30 MPa) and the external electric eld. When the sintering temperature rises from 950°C to 1050°C, much more elemental Sn is formed, and liquid Sn phase is migrated through Sn grains by viscous ow under external pressure. Next, liquid Sn grains are solidi ed in agglomerated form (Fig. 11) under a high cooling rate (~ 600°C/min) and high vacuum (1x10 − 5 bar).
Also, average grain size of SnO 2 is increased expectedly with the increment of temperature from 950°C to 1050°C that refers to grain growth governed by surface diffusion, where the size of isolated pores is not decreased through grain boundary and/or lattice diffusion that keeps densi cation rate constant [24].

Comparison of the sintering techniques
The relative density of SnO 2 ceramics remains constant at 63 %, although the sintering temperature increases from 1200°C to 1400°C during conventional sintering. SEM images of conventionally sintered SnO 2 ceramics are given in Fig. 12. It can be seen that the increasing sintering temperature causes a remarkable enhancement in grain size. The formation of elemental Sn from SnO 2 decomposition could not be observed.
The average size of equiaxed grains is found roughly 500 nm, 1 µm, and 3 µm for CS1-SnO 2 , CS2-SnO 2, and CS3-SnO 2 , respectively. Without any improvement in densi cation, the grain growth points out the surface diffusion occurred as a dominant mass transportation mechanism [24]. Therefore, 1200°C is chosen as an optimal sintering temperature for conventional sintering to prepare the sputtering target.
There is a signi cant difference in relative density values of SPS3-SnO 2 and CS-SnO 2 . The low relative density (63 %) indicates that the second step of sintering cannot be reached even at 1200°C for 120 min. The reason is related to not only lattice and grain boundary diffusion, but also plastic deformation and creep that took place during pressure-assisted sintering [24,26]. A higher density is achieved under using SPS at a lower temperature in reduced time because the densi cation rate relative to the coarsening rate increases by pulsed current (typically a few thousand amperes) under a pulsed DC voltage and external pressure as a capillary pressure. They promote mass transport by enhancing the total interfacial energy of curvature under deformation [18,24].
CM-SnO 2 has the same density value compared to CS-SnO 2 ceramics. It means that the commercial product was produced by a conventional (pressureless) technique.
XRD patterns given in Fig. 13 show that SPS3-SnO 2 , CS1-SnO 2, and CM-SnO 2 are composed of only SnO 2 phase. Figure 14 illustrates the SEM images (SE mode) of SPS3-SnO 2 , CS1-SnO 2 and CM-SnO 2 . The effect of the sintering technique on the microstructure of ceramics can be easily noticed. SPSed SnO 2 has a partially dense microstructure containing ne SnO 2 grains with an average size of ~ 300 nm.
Conventionally sintered SnO 2 has not a dense microstructure. Neck formation occurs without grain growth and pore isolation. SnO 2 grains in equiaxed shape have an average size of 500-600 nm. Like CS1-SnO 2, CM-SnO 2 possesses the same density and a non-dense microstructure where the sintering is limited with neck formation. However, the grain size is larger than 3 µm, and the grains are distributed inhomogeneously. That could be caused by a coarser starting powder of CM-SnO 2 compared to CS1-SnO 2 .

Conclusions
In the present study, SnO 2 ceramics were produced from undoped submicron SnO 2 powders using the SPS and pressureless-assisted sintering techniques by optimizing sintering conditions. The pressureless sintered SnO 2 ceramics at 1200°C-1400°C for 120 min exhibit very poor densi cation behavior and hence relatively low relative density (i.e., 63 % relative density) due to the dominance of non-densifying evaporation-condensation mechanism. On the other hand, monolithic SnO 2 ceramics were successfully produced by SPS at 950°C for 5 min, which demonstrated 93 % relative density and homogeneous grain size distribution. These results show that the pressure assistance in SPS dominates the densifying mechanisms against to non-densifying mechanisms. In addition, they reveal the effect of process conditions and sintering technique (sintering mechanism) on densi cation behavior, phase evolution, and microstructural development of SnO 2 ceramics. Accordingly, SPSed SnO 2 ceramics can be used as sputtering targets to produce high quality gas sensors due to their relatively high density (93 % relative density), high chemical purity, homogeneous grain size distribution and smaller average grain size. Rietveld analysis pattern obtained from powder diffraction data of SPS5-SnO2. The solid red lines are calculated intensities, and the black marked ones are observed intensities. The difference between the observed and calculated intensities is plotted below the pro le.   The temperature and displacement curves obtained through spark plasma sintering of a) SPS4-SnO2 and b) SPS5-SnO2 The schematic illustration of microstructural development of SPSed SnO2 ceramics as a function of time and temperature Figure 13 XRD patterns of SPS3-SnO2, CS1-SnO2 and CM-SnO2