Figure 8. Volume pore size distributions of mannitol powder tablets compressed with 72 MPa and stored in different moisture conditions 2 .
Due to electrostatic effects of mannitol powder in the sample cell, unfortunately, it was not possible to determine the effect of pretreatment on the pore structure of mannitol powder with mercury porosimetry. Mannitol powder adhered to the walls of the sample cell and came out of the sample cell during vacuum in the filling process. Due to swelling, after storage in moisture conditions and thus with increasing water content, the total pore volume of microcrystalline cellulose powder increased in low-pressure mercury porosimetry. According to the pore size results and the volume pore size distributions determined with low-pressure mercury porosimetry (i.e. pore diameter 14 – 200 m m), the volume of the smallest pores in microcrystalline cellulose powder was greatest when the samples were pretreated in vacuum conditions before measurement.
In the range of high-pressure porosimetry used in this work (pore diameter 7 nm – 14 m m), the total pore volume of microcrystalline cellulose powder decreased due to the moisture in the sample. Moisture did not affect other porosity parameters of the powder. Swelling increased the particle size of microcrystalline cellulose powder, and the voids between particles were determined in the range of low-pressure porosimetry. According to the theory on adsorption and condensation of water into pores, the smallest pores are filled first with water, which decreases the volume of the pores determined. However, no water-induced change in the volume of the smallest pores of microcrystalline cellulose was observed in high-pressure mercury porosimetry. According to the low- and high-pressure mercury porosimetry results of microcrystalline cellulose powder, proper pretreatment of the samples before mercury porosimetry analysis is important.
Pretreatment had no effect on the porosity parameters of mannitol granules in low-pressure porosimetry analysis. However, the total pore volume of microcrystalline cellulose granules increased due to swelling after storage in moisture conditions at the pore size range of low-pressure porosimetry. The median pore diameter of microcrystalline cellulose granules was smallest after storage in vacuum oven, which was evident also in the volume pore size distributions. The total pore surface area and volume of the smallest pores of microcrystalline cellulose granules increased with increasing moisture in low-pressure porosimetry analysis.
Moisture had no effect on the total pore volume of mannitol granules in high-pressure porosimetry analysis. However, the total pore volume of microcrystalline cellulose granules decreased due to the swelling with increasing moisture. With increasing moisture content, the total pore surface area of mannitol granules increased and the mean pore size decreased. The increase in the volume of the smallest pores of mannitol granules is also shown from the volume pore size distributions. The decrease in mean pore size was detected also in the microcrystalline cellulose granules.
During the adsorption, water fills the smallest pores determined with high-pressure mercury porosimetry first and the volume of these pores is supposed to decrease. Hearn and Hooton (1992) have presented that water would fill the pores of the samples and thus hinder the intrusion of mercury. Similarly, Ek et al. (1995) have suggested that mercury cannot intrude into the pores filled with another liquid. However, the volume of the smallest pores of the granules manufactured from mannitol and microcrystalline cellulose increased in this study. Because the water affects the volume of the smallest pores, it can be assumed that water settles into the smallest pores of granules. The structures of these granulated masses differ; mannitol granule mass consists of porous mannitol granules and partly of non-porous mannitol powder, whereas microcrystalline cellulose granules have a dense, non-porous structure. Water contents of the samples also differ remarkably. However, the increase in the volume of the smallest pores is most likely related to the complicated structure of granulated masses, because it was not observed with microcrystalline cellulose powder. One explanation could be that water is pushed through the pore structure of granules into new small pores in the face of an advancing mercury interface. On the other hand, Webb and Orr (1997) have suggested that the volume of the smallest pores and surface area values can be falsely large due to the compression of the samples during analysis. Thus, water may induce some compression of granulated samples during mercury porosimetry analysis. The median pore size was not affected, because this parameter reflects differences in the larger pore diameter range.
Pretreatment affects more the mercury porosimetry analysis of granules manufactured from hygroscopic material than of non-hygroscopic material, as expected. Pretreatment affected even the porosity parameters of non-hygroscopic mannitol granules, although the water contents of the samples were 1.2 % at the highest. Similar pretreatment of parallel samples before mercury porosimetry measurements is recommended.
Total pore volume of microcrystalline cellulose tablets compressed from powder increased due to the swelling of microcrystalline cellulose with increasing moisture, while the pore volume of mannitol powder tablets was unaffected. The change in microcrystalline cellulose tablets was observed after storage in 88% humidity. Moisture did not affect the volume of the smallest pores of mannitol tablets (pore diameter < 30 nm). However, for microcrystalline cellulose, mean pore size of tablets increased with increasing moisture. The maximum of the volume pore size distribution of mannitol tablets (pore size range 50 – 2000 nm, Fig. 8) and microcrystalline cellulose tablets (pore size range 500 – 2000 nm) changed towards larger pores with increasing humidity. This was observed also as increased median pore size values.
|
Figure 8. Volume pore size distributions of mannitol powder tablets compressed with 72 MPa and stored in different moisture conditions 2 . |
In tableting, the powder is bound together and the structure of the mass is densified. Therefore, water does not affect the structure of the tablets in the same way as it affects the starting materials. The possibility of capillary condensation increases due to the densified structure of the samples after tableting (El-Sabaawi & Pei 1977). In the present work, volume of the smallest determined pores of microcrystalline cellulose powder tablets decreased due to the water, whereas that of mannitol powder tablets remained unchanged. The water content in microcrystalline cellulose tablets is remarkably greater than that in mannitol tablets. The result of this work is consistent with the presentation of Hearn and Hooton (1992), that water in the sample behaves as a solid and thus hinders the intrusion of mercury. Similarly, according to Ek et al. (1995), mercury cannot intrude into pores already filled with another liquid.
Surprisingly, water settles in the pore size range of 50 - 2000 nm to the mannitol tablets and in the pore size range 500 – 2000 nm to the microcrystalline cellulose tablets. Some possible explanations are presented in the following. At the beginning of mercury porosimetry analysis, the sample is dried in order to fill the sample cell with mercury. Part of the water, especially from the smallest detected pores, is removed during this drying, and can be moved to the larger pores. During mercury porosimetry, the water on the surface of the sample can cause changes in the structure of the material studied under high pressure in the sample cell. The sample may for example be compressed during measurement. The water on the surface of the sample can be mobile under different conditions, and even promote chemical degradation or other types of physical changes (Ahlneck & Zografi 1990). On the other hand, puddles can be formed in the irregular pores of this size, which would decrease the volume of the pores (Allen et al. 1998). In addition to this, cyclohexane has been reported to cause a change in the contact angle of mercury on the surface of aluminium, and increase the determined pore size (Moscou & Lub 1981). Similarly, water on the surface of the tablet can change the contact angle of mercury and change the maximum of the pore size distribution towards larger pores.
According to the present results, the effect of pretreatment of samples appears to be very important when comparing tablets manufactured with direct compression. The pore structures of microcrystalline cellulose powder, granule and tablet samples are affected differently when stored in moisture conditions. The effect of pretreatment and water is observed even at the pore size range of larger pores (50 – 2000 nm) of tablets. Thus, if the effect of manufacturing on the pore structure of the sample is of interest, pretreatment of the samples should be similar before measurement.
The effect of water in the structure of tablets compressed from granules is even more complicated than its effects in the tablets manufactured by direct compression. Total pore volume of microcrystalline cellulose granule tablets increased due to the swelling with increasing moisture content, while the total pore volume of mannitol tablets was unaffected. The specific surface area of microcrystalline cellulose decreases and the structure is densified remarkably during wet granulation. The interaction and bonding between water and microcrystalline cellulose in humid conditions, which leads to swelling of microcrystalline cellulose, would appear to be different in the case of tablets compressed from granules when compared to tablets compressed from powder. However, according to moisture contents and total pore volume results of tablets, granulated mass adsorbs water and swells to the same extent after storage in humid conditions as does the powder in tablets. The water is thus evidently not adsorbed only to the outer surface of the granulated mass. Chatrath (1992) has also observed that microcrystalline cellulose granules adsorb water similarly to powder. According to her, hydrogen bonds formed into the granules during wet granulation break during adsorption of water into the granules. This theory appears to be correct according to this work.
Water did not decrease the volume of the smallest pores of mannitol tablets. However, volume of the smallest pores of microcrystalline cellulose granule tablets compressed with the highest compression pressure (196 MPa) increased with increasing moisture, as did that of the granules. Volume of the smallest pores of granule tablets compressed with lower compression pressures remained unchanged. Structure of hard microcrystalline cellulose granules is deformed when compressed with the highest compression pressure (196 MPa), which explains the result. Water molecules probably move in the structure of these granule tablets in front of the mercury that is intruding into the sample. Water can also cause some compression of these samples during a mercury porosimetry run. The median pore size of the tablets compressed from granules increased and the maximum of pore size distribution changed towards larger pores with increasing moisture (pore diameter range 50 – 2000 nm for mannitol tablets and 500 – 2000 nm for microcrystalline cellulose tablets), similarly to the tablets manufactured by direct compression. Although microcrystalline cellulose swells with increasing moisture, this trend is not related to the swelling, but to the settlement of the water molecules into the structure of tablets, to the maximum of the pore size range.
The effect of scanning speed on the pore structure of mannitol powder could not be determined. Mannitol adhered to the walls of the sample cell, and came out of the sample cell during filling with mercury. Scanning speed had no effect on the result of microcrystalline cellulose powder in low-pressure porosimetry analysis. Scanning speeds used in low-pressure analysis are low, and the differences between possible scanning speeds are small, which explains why no differences in low-pressure analysis were found.
Scanning speed had no effect on the total pore volume of microcrystalline cellulose powder in high-pressure mercury porosimetry. However, according to volume pore size distributions, the volume of the smallest pores of microcrystalline cellulose powder decreased with increasing scanning speed. This was shown also as decreased total pore surface area and increased mean pore size values with increasing scanning speed. According to the result, small pores of the powder were not accurately detected with fast scanning. Apparently, mercury does not have enough time to intrude into the smallest pores with fast scanning. Moscou and Lub (1981) have suggested a similar explanation. In the study of Hearn and Hooton (1992) on cement samples, likewise, scanning speed had no effect on total pore volume, but did have an effect on volume pore size distributions. The effect of mercury porosimetry scanning speed on the result of pharmaceutical samples has not been reported previously.
Scanning speed did not affect porosity values of mannitol or microcrystalline cellulose granules in low-pressure porosimetry analysis. In high-pressure porosimetry, scanning speed did not affect the total pore volume of mannitol or microcrystalline cellulose granules. Total pore surface area values were greatest and the mean pore size values smallest with the lowest scanning speed. Thus, the smallest pores of the granules were determined more accurately with the slowest scanning speed, which was evident also in the volume pore size distributions. No clear effect on the median pore size was observed, because this parameter emphasizes differences in the larger pore diameter range. The suggestion by Moscou and Lub (1981) that mercury may not have enough time to intrude into the pores is in accordance with the result of this study. Also, the result of Hearn and Hooton (1992), that scanning speed does not affect total pore volume values but volume pore size distributions, is consistent with the result of granules.
Scanning speed did not affect the total pore volume values of mannitol or microcrystalline cellulose tablets manufactured by direct compression. However, the smallest mean pore size was observed with the slowest scanning speed, which was evident also in the volume pore size distribution curves as the greatest volume of the smallest pores. Thus, with the fastest scanning the volume of the smallest pores is lowest because of the lack of time for the mercury to intrude into the pores properly.
The maximum of the volume pore size distribution of mannitol powder tablets (pore diameter 1000 nm) and microcrystalline cellulose powder tablets shifted towards smaller pore sizes (pore size range 100 – 1000 nm) with increasing scanning speed. The median pore size, which emphasizes changes at this large pore size range, was unaffected for mannitol tablets, whereas the median pore size of microcrystalline cellulose tablets decreased with increasing scanning speed. This result is consistent with the shift of maximum of pore size distribution towards smaller pore sizes. The result is related to the structure of tablets, because it was not observed in the measurements of powders or granules. This is probably because with higher scanning speeds mercury does not have time to intrude into the pores of this size range at the right time. Intrusion takes place later and the intruded mercury is detected at the smaller pore size range. The pore structure of direct compressed tablets is more rigid than that of powder and granules. No packing or rearrangement of individual particles, which is possible in mercury porosimetry measurement of powder and granules, takes place during intrusion of mercury into the tablets.
The total pore volumes of mannitol and microcrystalline cellulose tablets compressed from granules were unaffected by scanning speed. The mean pore size of tablets compressed from mannitol and microcrystalline cellulose granules with the smallest compression pressure was smallest with the slowest scanning speed. Thus, the smallest pores of granule tablets are also determined more accurately with slow scanning. This result can be observed also from the volume pore size distributions. However, the mean pore size of microcrystalline cellulose tablets compressed from granules with the two highest compression pressures (122 and 196 MPa) was unaffected by the scanning speed.
Surprisingly, in contrast with the result of mannitol tablets manufactured by direct compression, the median pore size of mannitol granule tablets increased with increasing scanning speed. The median pore size values of granule tablets are lower than those of powder tablets. This denser structure together with the more complex pore structure of granule tablets are the reasons why the effect of scanning speed is different in granule tablets. However, the shift of determined pore size was so small that it was not observed in the volume pore size distributions. In contrast to the result of mannitol granule tablets, the median pore size of microcrystalline cellulose granule tablets decreased with increasing scanning speed. Similarly to tablets manufactured with direct compression, the maximum in the pore size range 100 – 1000 nm changed towards smaller pores with increasing scanning speed. The pore structure of granule tablets is more complicated than that of powder tablets, and thus the effect of scanning speed is not similar. Based on these results, no clear conclusions can be drawn on the effect of scanning speed on the pore structure of tablets compressed from granules.
Pores in mannitol powder i.e. voids between particles, were determined with high-pressure porosimetry in the diameter range 1 – 5 m m. The volume of these pores decreased markedly during wet granulation, and new intragranular pores were formed in the diameter range 40 – 300 nm. This was observed with both high-pressure porosimetry and nitrogen adsorption. These intragranular pores were formed when powder particles were dissolved during granulation and recrystallised on the larger particles (Juppo 1995). According to Juppo (1995), small particles also attach to each other by solid bridges formed by recrystallised mannitol or by binder. According to mercury porosimetry and nitrogen adsorption, granules were more porous than powder.
Densification of the powder mass with increasing compression pressure was detected in the pore diameter range from 7 nm to 14 m m from the total pore volume and pore size values obtained by mercury porosimetry. From the volume pore size distribution curves of powder and tablets compressed from powder measured with mercury porosimetry, densification was observed in the pore diameter range from 200 to 2000 nm (Fig. 9(A)). These pores are the voids between powder particles. The largest pores disappeared first, pore size decreased, and the maximum of the distribution moved towards the smaller pores indicating plastic deformation. However, a new pore population in the pore size range from 20 to 50 nm was created in tablets compressed with the highest compression pressure (196 MPa) due to the fragmentation of powder (Fig. 9(A)). Fragmentation increased the number of small particles, contributing to the appearance of a new group of pores (Vromans et al. 1985). This new pore population was related to increased breaking force of the tablets, which was almost similar for tablets compressed at the two lowest compression pressures, 72 and 122 MPa. When the number of pores larger than 500 nm decreases and the number of pores smaller than 200 nm increases, breaking force of tablets increases (Juppo 1996c).
With nitrogen adsorption, the pore volume of mannitol powder in the pore diameter range from 3 to 200 nm increased when compressed, indicating formation of new pores in the pore size range measured. Size of the voids between powder particles decreased, and these voids were determined at the detection range of nitrogen adsorption. No difference in pore volume of tablets compressed from powder with different compression pressures was observed. The pore size distribution obtained by nitrogen adsorption had only one maximum for tablets compressed at the two lowest pressures (72 and 122 MPa). Bimodal distribution was created after compression at the highest pressure, 196 MPa, indicating fragmentation of powder particles. The specific surface area of tablets determined with nitrogen adsorption increased with increasing compression pressure, also indicating slight fragmentation of mannitol at this pore size range (3 – 200 nm).
Deformation of granules is observed from the total pore volume, mean and median pore size and the volume pore size distributions of granule tablets determined with mercury porosimetry; the largest pores of tablets disappeared with increasing compression pressure due to fragmentation of granules (Fig. 9(B)). Juppo (1996b) has reported fragmentation of granules with increasing compression pressure when measured with mercury porosimetry. In the present work, pores of mannitol granules were unaffected by the lowest compression pressure. When higher compression pressures were used, deformation shifted the maximum of pore size distribution to smaller values. Due to the fragmentation of granules, more small pores (diameter less than 20 nm) were created in the tablets compressed with the two highest compression forces, 122 and 196 MPa (Fig. 9(B)). The broad size distribution of mannitol granules is still detectable in tablets. Selkirk and Ganderton (1970) have shown that granules have caused a wider pore size distribution for tablets than powder, which is consistent with the result of mannitol tablets in this study.
The volume pore size distribution of granule tablets measured with nitrogen adsorption is bimodal, one maximum showing pores of the granules in the pore size range from 50 to100 nm. The volume of these intragranular pores is highest in the tablets compressed at the smallest compression pressure (72 MPa). Due to densification, the volume of these pores decreases with increasing compression pressure. The volume of the smallest detectable pores (diameter < 7 nm) increases with increasing compression pressure, probably due to fragmentation. Specific surface area of granule tablets decreased with increasing compression pressure due to plastic deformation of the mass at this pore size range (3 – 200 nm). Thus, plastic deformation and fragmentation of mannitol granules were observed with nitrogen adsorption.
|
(A) (B) Fig. 9. Volume pore size distributions determined with mercury porosimetry (A) a) mannitol powder and mannitol powder tablets compressed with b) 72 MPa c) 122 MPa and d) 196 MPa. (B) a) mannitol granules and mannitol granule tablets compressed with b) 72 MPa, c) 122 MPa and d) 196 MPa 3 . |
The breaking forces of the granule tablets were markedly higher than those of powder tablets, indicating that wet granulation improves the compactibility of mannitol. During compression, greater densification of the granules was observed at the detection range of high-pressure porosimetry (i.e. pore diameter 7 nm – 14 m m) when compared to the densification of the powder mass. Good compressibility and strong tablets compressed from mannitol granules has been reported by Juppo et al. (1995). Larger specific surface area values of granules and of tablets compressed from granules, obtained with mercury porosimetry and nitrogen adsorption, than of powder and of tablets manufactured by direct compression also indicate a more porous structure and deformation of granules during compression. Under compression, large porous granules are deformed more than is needle-shaped mannitol powder. Mannitol powder deforms more plastically (Roberts & Rowe 1985), and mannitol granules by fragmentation and plastic deformation (Juppo et al. 1995). Plastic deformation of mannitol powder was observed from the porosity parameters obtained with mercury porosimetry. However, some fragmentation of also mannitol powder takes place, as observed from the volume pore size distributions obtained with mercury porosimetry and nitrogen adsorption, and from the specific surface area results determined with nitrogen adsorption. Fragmentation of granules was observed from the volume pore size distributions obtained with both methods.
The specific surface areas obtained with nitrogen adsorption and the porosity parameters determined with mercury porosimetry show plastic deformation of mannitol granules. According to Krycer et al. (1982), the crushing strength of tablets compressed from mannitol powder or granules increases with increasing porosity of raw material. The porosity percentages and total pore volumes of granule tablets decrease more under compression due to greater deformation than those of powder tablets, when tablets compressed with 72 MPa and 196 MPa are compared. In general, tablets compressed from granules have higher strength when compared to those compressed from powder. Consistent with the result of mannitol in this study, strength is related to the large area available for bond formation and to the material undergoing fragmentation (Nyström et al. 1993).
The structure of microcrystalline cellulose was densified in wet granulation. Slight densification of the powder mass after wet granulation was observed from the volume pore size distribution obtained with mercury porosimetry. New pores were not formed, which generally takes place during granulation. Millili (1990) has explained densification during pelletisation by autohesion, which is not related to hydrogen bonding. According to Kleinebudde (1997), hydrogen bonds are formed between crystallites or their agglomerates during pelletisation and drying (crystallite gel model). However, wet granulation and pelletisation are not directly comparable processes, and microcrystalline cellulose behaves differently during these two processes. Chatrath (1992) has explained her theory of increased hydrogen bonding in wet granulation by a similar ability of microcrystalline cellulose powder and granules to adsorb water vapor. In her study, intraparticular bonds in granules were more disrupted during water vapor adsorption than those in microcrystalline cellulose powder, which explains the hydrogen bonding theory. Similarly in our study, water vapor adsorption in powder and granules was equal, and thus densification in wet granulation is related to hydrogen bonding. Buckton et al. (1999) have observed the increased intraparticular hydrogen bonding of microcrystalline cellulose after wet granulation with near IR technique. Densification of microcrystalline cellulose in wet granulation according to this work takes place at the determination range of nitrogen adsorption, in the pore diameter range from 3 to 200 nm.
The structure of the granules was so dense and the volume of the pores so small that the pore volume or the volume pore size distribution could not be determined with the Coulter SA 3100 nitrogen adsorption method. However, the specific surface area of microcrystalline cellulose obtained with nitrogen adsorption decreased markedly after wet granulation, indicating densification of the mass. A similar result has been obtained by Chatrath (1992).
Deformation of powder and decrease in the size of the voids between powder particles after compression was observed from the volume pore size distribution obtained with mercury porosimetry (Fig. 10(A)). The maximum in the pore diameter range from 200 to 2000 nm shifted towards smaller pores, and the volume of pores smaller than 40 nm decreased. Consistent with this study, Vromans et al. (1985) have reported that the volume pore size distribution of tablets compressed from microcrystalline cellulose powder shifted to smaller pore diameters with increasing compression force. In this study, with increasing compression pressure, the total pore volume and mean and median pore size values decreased, indicating densification of the mass. In a study by Sixmith (1977), the modal pore radius of microcrystalline cellulose tablets compressed from powder decreased with increasing compression pressure. According to the volume pore size distribution obtained with mercury porosimetry in this study, microcrystalline cellulose deforms plastically, and no evidence of fragmentation was found. Microcrystalline cellulose is known as a material that deforms plastically (Lamberson & Raynor 1976, David & Augsburger 1977, Schangraw et al. 1981, Staniforth et al. 1988). Hydrogen bonding and mechanical interlocking of irregular particles together with a large particle surface area and filamentous structure of microcrystalline cellulose lead to good compressibility of powder (Bolhuis & Lerk 1973). In this study, the total pore surface area of powder determined with mercury porosimetry decreased when compressed. However, the total pore surface area of tablets does not change with increasing compression pressure.
Unexpectedly, the total pore volume of powder determined with nitrogen adsorption in the pore diameter range from 3 to 200 nm was greater when compressed with 122 MPa and 196 MPa when compared to values of powder and powder tablets compressed with 72 MPa. Sixmith (1977) has reported an increased surface area of Avicel Ò tablets when compression pressure exceeded 125 MPa. According to volume pore size distributions of tablets obtained with nitrogen adsorption, the volume of the pores decreased with increasing compression pressure, indicating plastic deformation of microcrystalline cellulose in this pore size range. The pore volume is determined in the adsorption phase, while the volume pore size distribution is measured from the desorption phase. The reason for the pore volume result may be the opening up of closed pores of microcrystalline cellulose in compression (Sixmith 1977). The specific surface area of powder tablets measured with nitrogen adsorption decreased with increasing compression pressure due to the plastic deformation of microcrystalline cellulose in compression.
Deformation of granules was observed from the volume pore size distribution curves in the pore diameter range from 500 to 2000 nm as a shift of maximum towards smaller pores (Fig. 10(B)). The decrease in the volume of pores < 50 nm in diameter is clearly observed when pore volumes of granule tablets compressed with 72 and 122 MPa are compared with those of tablets compressed with 196 MPa. Therefore, the mean pore size increased and total pore surface area decreased between compression pressures 122 and 196 MPa. This change is in agreement with the increased breaking force values of granule tablets when compression pressure exceeds 122 MPa. Johansson et al. (1998), similarly, reported an increase in tensile strength of microcrystalline cellulose tablets compressed from pellets when compression pressure reached as high as 160 MPa. According to Staniforth et al. (1988), most of the compression force was used to break up the primary granule structure of microcrystalline cellulose. Schwartz et al. (1994) have observed some fracture and plastic deformation of microcrystalline cellulose pellets during compression. According to Maganti and Celik (1993), the bonding of microcrystalline cellulose decreased in pelletisation due to changes in shape and size and the reduction of bonding sites after pelletisation. They reported elastic deformation and brittle fragmentation of microcrystalline cellulose pellets in compression. Deformation, densification and only limited fragmentation of microcrystalline cellulose pellets has occurred in compression (Johansson et al. 1995, Johansson & Alderborn 1996, Johansson et al. 1998).
The specific surface area of granule tablets decreased with increasing compression pressure when determined with nitrogen adsorption (pore diameter range 3 – 200 nm). However, an increase was observed in specific surface area when granules were compressed with 72 MPa due to fragmentation. Unfortunately, the structure of the granules was so dense and thus the volume of the pores so small that other porosity parameters could not be determined with the Coulter SA 3100 nitrogen adsorption method.
|
(A)
(B) Fig. 10. Volume pore size distributions determined with mercury porosimetry (A) a) microcrystalline cellulose powder and microcrystalline cellulose powder tablets compressed with b) 72 MPa c) 122 MPa and d) 196 MPa. (B) a) microcrystalline cellulose granules and microcrystalline cellulose granule tablets compressed with b) 72 MPa, c) 122 MPa and d) 196 MPa 4 . |
The breaking forces of microcrystalline cellulose tablets compressed from granules were markedly lower than those of the tablets compressed from powder. Similarly, Staniforth et al. (1988) observed greater strength of interparticle bonding for powder samples of microcrystalline cellulose than for granulated mass. In this study, the porosity percent of powder decreased more when compressed than that of granules. This indicates greater densification of microcrystalline cellulose powder in compression. Thus, the compressibility of microcrystalline cellulose decreased in wet granulation. This result is consistent with the result of the study of Staniforth et al. (1988). Microcrystalline cellulose powder deforms plastically (Lamberson & Raynor 1976, David & Augsburger 1977, Shangraw et al. 1981, Staniforth et al. 1988), whereas deformation in combination with densification and only limited fragmentation of pellets manufactured from microcrystalline cellulose has been observed (Johansson et al. 1995, Johansson & Alderborn 1996, Johansson et al. 1998). Plastic deformation of microcrystalline cellulose powder is clear from the results obtained in this study with mercury porosimetry and nitrogen adsorption. For powder, no evidence of fragmentation was observed. The structure of granules was deformed when compression pressure reached 196 MPa, because the breaking force of tablets increased remarkably between the compression pressures of 122 and 196 MPa. Due to deformation, the volume of the pores < 50 nm in diameter decreases between the compression pressures of 122 and 196 MPa. Similarly to the result of this work, Johansson et al. (1998) have reported increased tensile strength of tablets compressed from microcrystalline cellulose pellets when compression pressure reached 160 MPa. However, in the present work, densification of granules was observed in the detection range of nitrogen adsorption (pore diameter range 3 – 200 nm) as decreasing specific surface area values with increasing compression pressure. In this study, decreased compactibility of microcrystalline cellulose after wet granulation was related to the smaller specific surface area values of granules when compared to those of powder.
Webb and Orr (1997) have suggested that pore structures obtained with mercury porosimetry and nitrogen adsorption are comparable only if the pore size range from 3 to 300 nm is compared. In this work, volume pore size distributions are compared in the overlapping pore size range (i.e. diameter range from 7 nm to 200 nm). However, the total pore volume, surface area and volume pore size distributions obtained with these methods are compared as they are obtained with these methods without applying corrections. 463188466"> One aim of this work was to study how to use mercury porosimetry and nitrogen adsorption in an effective and correct way in the analyses of pharmaceutical samples. In the pharmaceutical industry, the use of these methods is easier if the results can be evaluated as they are determined. That is why corrections were not applied to the results. Volume pore size distributions were determined with mercury porosimetry from the intrusion phase and with gas adsorption from the desorption phase, because the obtained distributions describe the pore structure similarly (Conner et al. 1986).
Due to different measurement ranges, nitrogen adsorption gave markedly smaller total pore volume values for mannitol and microcrystalline cellulose powders than did mercury porosimetry. According to Webb and Orr (1997), the pore volume measured with mercury porosimetry is larger than the one determined with nitrogen adsorption if the sample contains pores larger than 300 nm. Pharmaceutical powders tend to have low porosity in the detection range of nitrogen adsorption. Mercury porosimetry determines also the voids between the particles, which affect the determined volume more than does the internal porosity of the particles. Milburn et al. (1991) have obtained similar pore volume values for silica samples with these methods. Pores of the silica samples, however, were markedly smaller and the structure of silica more porous than those of mannitol and microcrystalline cellulose powders. Thus, the pores were determined mainly in the detection range of nitrogen adsorption. Stanley-Wood (1978) and Conner et al. (1986) have obtained almost the same pore size distributions for magnesium trisilicate and for Degussa aerosols with these techniques. However, non-similar pore size distributions have been obtained for silicas, iron oxide-chromium oxide catalyst, aerosil powder and chrysotile powder (Brown & Lard 1974, DeWit & Scholten 1975). Moscou and Lub (1981) and Johnston et al. (1990) have reported damage or compression of highly porous silica and aluminium samples during mercury porosimetry. In this study, pores in the mannitol and microcrystalline cellulose powders were detected in the same pore size range with both methods in the overlapping pore size area, which indicates that no compression of the samples takes place during mercury porosimetry. Although the pore size distributions had a similar shape, the intensities of the curves were different.
Total pore volume of mannitol granules determined with mercury porosimetry was markedly larger than that obtained with nitrogen adsorption, as it was for mannitol and microcrystalline cellulose powder. The structure of microcrystalline cellulose granules was so dense, unfortunately, that the pore structure could not be determined with Coulter SA 3100 nitrogen adsorption method. The volume of the pores was so small that it was out of the detection range of the method. However, pores in the mannitol granules were detected in the same pore size range with both methods. The intensities of the curves were different for the two methods, as they were for powders. The structure of the mannitol granules was not destroyed or compressed during mercury porosimetry.
The total pore volumes of mannitol and microcrystalline cellulose tablets determined with mercury porosimetry were markedly higher than those measured with nitrogen adsorption, as they were for mannitol powder and granules and microcrystalline cellulose powder. Mercury porosimetry determines larger pores that are not within the detection range of nitrogen adsorption and which have more effect on the total volume.
The volume pore size distributions of mannitol tablets measured with nitrogen adsorption and mercury porosimetry had the same shape in the overlapping pore size region, although the scales of the curves differed from each other. Damage or compression of highly porous particles such as silica and alumina samples has been reported (Moscou & Lub 1981, Johnston et al. 1990). Judged by the consistent pore size distributions obtained with both nitrogen adsorption and mercury porosimetry, no compression or damage of the mannitol tablets took place during mercury porosimetry analysis. However, the volume pore size distributions of microcrystalline cellulose tablets compressed from powder were not equal in the overlapping pore size range when determined with these methods. The microstructure of a microcrystalline cellulose tablet may be deformed in analysis. According to Webb & Orr (1997), compression of the samples in mercury porosimetry can be observed as a large volume of medium-sized or small pores. In this work, maximum of the pore size distribution determined with mercury porosimetry was in the smallest detectable pore size range (i.e. diameter < 10 nm). At this pore size range, no maximum was detected in distribution obtained with nitrogen adsorption. Faroongsarng and Peck (1994) have reported consistent pore size distributions of dicalcium phosphate dihydrate tablets obtained by nitrogen adsorption and mercury porosimetry in the overlapping pore size range. Also, Stanley-Wood (1978) has reported almost similar pore size distributions for magnesium trisilicate and Conner et al. (1986) for Degussa aerosols when determined with these techniques. However, similarly to the present result with microcrystalline cellulose tablets, Brown and Lard (1974) and De Wit and Scholten (1975) obtained non-similar pore size distributions for silicas, iron oxide-chromium oxide catalyst, aerosil powder and chrysotile powder. Differences were explained with compression of highly porous silica, non-capillary pore structure of samples and limitations of the Washburn equation in characterising the smallest detectable pores during mercury porosimetry. In our study, however, microcrystalline cellulose tablets remained whole after porosimetry measurement.
The surface area values of mannitol and microcrystalline cellulose powder, granules and tablets obtained with mercury porosimetry are markedly higher than those measured with nitrogen adsorption. This is because mercury porosimetry determines larger pores than nitrogen adsorption, and further because of the complex pore structure of the samples, ink-bottle shaped pores and low porosity of pharmaceutical samples. Surface area in mercury porosimetry is calculated from the volume intruded in pore diameter intervals, assuming cylindrical pores with a round pore opening. Ink-bottle pores tend to increase surface area values calculated from mercury porosimetry data, because the volume of pores with a small neck can be remarkable. Dees and Polderman (1981) have reported higher surface area values with mercury porosimetry than with nitrogen adsorption for lactose tablets. They concluded that nitrogen adsorption results were more accurate. A similar result has been obtained also with silica samples (Milburn et al. 1991). In contrast, Mikijelj and Varela (1991) have found the pore surface areas of magnesium oxide and diatomite compacts measured with these methods to be equivalent. In their study, the highest pressure in mercury porosimetry was 103 MPa, and the diameter of the smallest detectable pores 14 nm. Surface area values of the samples were 2 - 50 m 2 /g, indicating that the pores were very small. Thus, the pores were probably mainly in the detection range of nitrogen adsorption. Adkins and Davis (1988) have made the surface area values of alumina and zirkonia comparable by correcting the contact angle used in mercury porosimetry. In their study, surface areas of the samples were from 46 to 130 m 2 /g. With higher surface areas, results were no longer comparable. According to Milburn and Davis (1993), the correlation between surface areas obtained with these methods in samples of very low surface areas is poor. In the present study on pharmaceutical samples, no corrections were made to these parameters, and nitrogen adsorption was more capable of detecting changes in the tablet surface area caused by tableting.
2 Reprinted from Pharmaceutical Development and Technology, 5(2), Westermarck et al., Mercury porosimetry of mannitol tablets: effect of scanning speed and moisture, p. 186, copyright (2000), Marcel Dekker.
3 Reprinted from European Journal of Pharmaceutics and Biopharmaceutics, 46, Westermarck et al., Pore structure and surface area of mannitol powder, granules and tablets determined with mercury porosimetry and nitrogen adsorption, p. 66, copyright (1998), Elsevier Science.
4 Reprinted from European Journal of Pharmaceutics and Biopharmaceutics, 48, Westermarck et al. Microcrystalline cellulose and its microstructure in pharmaceutical processing, p. 204, copyright (1999), Elsevier Science.
