- Email: [email protected]
Unusual Effect of Water Vapor Pressure on Dehydration of Dibasic Calcium Phosphate Dihydrate
Unusual Effect of Water Vapor Pressure on Dehydration of Dibasic Calcium Phosphate Dihydrate ADITYA M. KAUSHAL,1 VENU R. VANGALA,1,2 RAJ SURYANARAYANAN1 1
Department of Pharmaceutics, College of Pharmacy, University of Minnesota, Minneapolis, Minnesota 55455
2 Venu R. Vangala’s present address is Institute of Chemical and Engineering Sciences, A∗STAR (Agency for Science, Technology and Research), Jurong Island, Singapore 627833
Received 17 August 2010; revised 8 September 2010; accepted 20 September 2010 Published online 6 December 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.22372 ABSTRACT: Dibasic calcium phosphate occurs as an anhydrate (DCPA; CaHPO4 ) and as a dihydrate (DCPD; CaHPO4 •2H2 O). Our objective was to investigate the unusual behavior of these phases. Dibasic calcium phosphate dihydrate was dehydrated in a (i) differential scanning calorimeter (DSC) in different pan configurations; (ii) variable-temperature X-ray diffractometer (XRD) at atmospheric and under reduced pressure, and in sealed capillaries; and (iii) water vapor sorption analyzer at varying temperature and humidity conditions. Dehydration was complete by 210◦ C in an open DSC pan and under atmospheric pressure in the XRD. Unlike “conventional” hydrates, the dehydration of DCPD was facilitated in the presence of water vapor. Variable-temperature XRD in a sealed capillary and DSC in a hermetic pan with pinhole caused complete dehydration by 100◦ C and 140◦ C, respectively. Under reduced pressure, conversion to the anhydrate was incomplete even at 300◦ C. The increase in dehydration rate with increase in water vapor pressure has been explained by the Smith–Topley effect. Under “dry” conditions, a coating of poorly crystalline product is believed to form on the surface of particles and act as a barrier to further dehydration. However, in the presence of water vapor, recrystallization occurs, creating cracks and channels and facilitating continued dehydration. © 2010 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 100:1456–1466, 2011 Keywords: dibasic calcium phosphate; dehydration; water vapor; calorimetry (DSC); thermogravimetric analysis; X-ray diffractometry; excipients; hydrate; physical stability
INTRODUCTION Solid oral dosage forms typically consist of an active pharmaceutical ingredient (API) and multiple excipients. The performance of the dosage form may be influenced by the physical forms of the formulation components. A large number of studies have focused on the physical form [the term “physical form” encompasses polymorphs, solvates, the noncrystalline (amorphous) form, and partially crystalline forms] of APIs, both during the manufacture of the dosage form and its subsequent storage. The implications of such phase transformations, both during pharmaceutical processing and subsequent storage, on product performance have also been evaluated.1,2 For example, a number of APIs are marketed as hydrates, wherein Correspondence to: Raj Suryanarayanan (Tel: 612-624-9626; Fax: 612-626-2125; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 100, 1456–1466 (2011) © 2010 Wiley-Liss, Inc. and the American Pharmacists Association
1456
water molecules are incorporated, usually stoichiometrically, in the crystal lattice. Dehydration during storage can have implications on product performance and this issue has received considerable attention.3,4 However, such in-depth attention has not been paid to the phase behavior of excipients. When the weight fraction of the API in the dosage form is low, excipients, especially diluents, constitute a major proportion of the dosage form. Phase transformations of excipients either during processing or during subsequent storage can be equally serious because they can also influence the dosage form performance. Dibasic calcium phosphate is widely used as a diluent in tablets and capsules, and occurs as dibasic calcium phosphate anhydrate (DCPA; CaHPO4 ) and as dibasic calcium phosphate dihydrate (DCPD; CaHPO4 •2H2 O).5–8 Both DCPD and DCPA exhibit good compressibility, amenability to scale-up, and brittle fracture propensity, making them popular excipients. Dibasic calcium phosphate dihydrate has
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1457
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
a stoichiometric water content of 20.9% (w/w). As a diluent, it can constitute a substantial fraction (over 50%, w/w) of the dosage form.9,10 The dehydration of DCPD to yield DCPA has been documented in numerous studies.11–14 Given its high water content, dehydration of even a fraction of DCPD in the formulation can lead to the release of an appreciable amount of water. The released water can cause chemical or physical changes in the dosage form components. Aspirin degradation during accelerated stability testing [35◦ C and 82.9% relative humidity (RH)] has been attributed to the water released by DCPD dehydration.10,15 In addition, the functionality of excipients can be affected by the released water. For example, in superdisintegrants, the sorption of the released water can cause product failure. Finally, changes in tablet hardness and dimensions were attributed to in situ DCPD dehydration.16,17 It is important to recognize that DCPD dehydration in situ may not be readily apparent, if the liberated water is taken up by one or more formulation components. There will not be any change in tablet weight because water is transferred from one formulation component to another. In other words, the effect is brought about by a change in the distribution of water in the dosage form. The dehydration behavior of DCPD alone18 was evaluated as the first step toward understanding the dehydration in complex multicomponent formulations. When heated in a thermogravimetric analyzer (TGA) under nitrogen purge, the dehydration was complete by 210◦ C. When DCPD was dehydrated at 40◦ C, but at RH values ranging from 2% to 75%, the dehydration kinetics was heavily influenced by the water vapor pressure. The most rapid dehydration, based on the observed weight loss, occurred at intermediate RH values of 32% and 52% RH. On the contrary, DCPA was remarkably resistant to hydration, and it was stable even when dispersed in water at 50◦ C for 7 months.18 A conventional anhydrate–hydrate pair, under isothermal conditions, can exist in stable equilibrium at the transition water vapor pressure (pt ), and the anhydrate is stable at water vapor pressures
in a TGA, differential scanning calorimeter (DSC), and in an X-ray diffractometer (XRD), it was possible to comprehensively characterize the reactant (DCPD) and the product (DCPA) phases.
EXPERIMENTAL Materials R
DCPD from a commercial source (Emcompress batch 7070X, JRS Pharma, Patterson, New York) was used for the work. Differential Scanning Calorimetry The sample (5–10 mg) was heated in a DSC (Model 2920, TA Instruments, New Castle, Delaware) from 25◦ C to 250◦ C at 10◦ C/min under nitrogen purge (75 mL/min). Two different pan configurations, open pan (no lid) and hermetically sealed pan with standard 75 :m pinhole, were used. The experimental data were analyzed using a commercially available software (TA Universal Analysis 2000; TA Instruments). Thermogravimetric Analysis TGA was employed in both isothermal and nonisothermal modes. For the nonisothermal analysis, 5–10 mg of the sample was heated in the TGA (Model Q50, TA Instruments) from room temperature to 250◦ C at 10◦ C/min under nitrogen purge (75 mL/min). Both the pan configurations, as in the DSC, were utilized. For the isothermal analyses, the temperature was rapidly increased to the desired value and maintained isothermally thereafter. Variable-Temperature X-Ray Diffractometry
Under Reduced Pressure Approximately 100 mg of sample was filled in an aluminum holder by the top-fill technique and exposed to CuK" radiation (45 kV × 40 mA) in a wide-angle XRD (model XDS 2000, Scintag Inc., Cupertino, California). The sample holder was covered with a stainless steel dome having a beryllium window and evacuated R to ∼50 mTorr. A temperature controller (Micristar , Model 828D; TMC Services, Inc., Elk River, Minnesota) was used to heat the sample from 25◦ C to 300◦ C. The angular range was 5◦ to 40◦ 22 and counts were accumulated for 1 s at each step. During the XRD run (∼12 min), the sample was maintained under isothermal conditions. Data analyses were performed using commercially available software (JADE; Materials Data, Inc., Livermore, California).
In Sealed Capillaries DCPD was sealed in a glass capillary (0.7 mm diameter; Charles Supper Company, Natick, Massachusetts) and exposed to CuK" radiation JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1458
KAUSHAL, VANGALA, AND SURYANARAYANAN
(50 kV × 100 mA; collimated to 0.5 mm i.d. spot size) generated with a rotating anode (18 kW; Rotaflex generator, Woodlands, Texas). A two-dimensional area detector (Highstar, D8 Discover Platform; Bruker AXS, Madison, Wisconsin) was used at a fixed position of 22◦ 22, which covered the angular range from 7◦ to 37◦ 22. Capillaries were heated from 25◦ C to 100◦ C using a temperature controller (3000CN, Omega Engineering, Stamford, Connecticut).
tion of hydrates.19,22 However, the opposite effect was observed here indicating that dehydration of DCPD was facilitated in the presence of water vapor. The enthalpy of transition (dehydration accompanied by water vaporization) was nearly identical in the two pan configurations: 567.5 ± 5.4 (mean ± standard deviation; n = 4) in open pans and 570.9 ± 4.7 (mean ± standard deviation; n = 4) in hermetic pans with pinhole.
Dehydration in an Automated Water Sorption/Desorption Apparatus
Variable-Temperature XRD
Approximately 20 mg of the sample was placed in the quartz sample pan of an automated vapor sorption analyzer (TGA Q5000SA; TA Instruments) and equilibrated at 60◦ C/0% RH. The RH was then rapidly changed to a fixed value (ranging from 0% to 80%, 60◦ C), where the sample was held for 1800 min and the weight was continuously monitored.
RESULTS AND DISCUSSION The dehydration behavior of DCPD was evaluated under (i) nonisothermal and (ii) isothermal conditions. Dehydration Under Nonisothermal Conditions
Differential Scanning Calorimetry and Thermogravimetric Analysis The DSC curve of DCPD, when heated in an open pan, was substantially similar to the profile reported earlier.18 The three endotherms were attributed to dehydration accompanied by vaporization of water (Fig. 1a). This conclusion was based on the fact that in the temperature range where the endotherms were observed in the DSC, a weight loss was observed in the TGA. The derivative TGA plots enabled us to discern the precise temperature range of the weight loss. Dehydration appeared to be complete by ∼210◦ C. The thermal behavior of DCPD, characterized by DSC, TGA, and variable-temperature XRD, was discussed in our earlier publication.18 In our previous DSC studies, only open pans (without lids) were used. It is well known that the dehydration behavior of hydrates can be influenced by the DSC pan configuration.19–21 In a hermetically sealed pan with a pinhole, the released water will not leave the pan until the water vapor pressure in the pan is substantially greater than the atmospheric pressure. Therefore, the solid will experience an elevated, though uncontrolled water vapor pressure. In this configuration, the dehydration of DCPD appeared to be a single-step process and was complete by ∼140◦ C (Fig. 1b). The weight loss observed in a TGA, also in a hermetically sealed pan with a pinhole, enabled us to confirm these results. Conventionally, an increase in water vapor pressure will decelerate the dehydraJOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
In an effort to obtain phase information, DCPD was subjected to variable-temperature XRD under ambient pressure. As reported previously, DCPD directly and completely dehydrated to DCPA by ∼220◦ C with no evidence of any intermediate phases.18 Even at a rapid heating rate of 40◦ C/min, no halos, attributable to amorphous phases, were seen. Because we could not simulate the exact DSC pan configuration (sealed pan with a pinhole), DCPD was subjected to variabletemperature XRD in a sealed capillary. When heated in a sealed capillary, dehydration appeared to be complete by ∼100◦ C (disappearance of the 21.0◦ 22peak of DCPD) with the concomitant appearance of the DCPA peaks at 26.4◦ and 30.2◦ 22 (Fig. 2b). In this configuration, the water released by dehydration can be in intimate contact with the solid and this appeared to substantially accelerate the dehydration of DCPD. In an effort to confirm the role of water vapor, DCPD was heated under reduced pressure (∼50 mTorr). Under this condition, the water vapor released upon dehydration would be quickly removed, thereby limiting the interaction between water vapor and solid. Interestingly, the dehydration of DCPD was incomplete even at 300◦ C (Fig. 2b). Although the characteristic peak of DCPA at 26.4◦ 22 increased in intensity with temperature, the DCPD peaks (21.0◦ and 11.7◦◦ 22) remained detectable through the duration of the experiment. To verify that the dehydration was incomplete at 300◦ C, the solid was cooled to room temperature and subjected to TGA and DSC in an open pan. In the TGA, a sharp weight loss of 11.6% at ∼175◦ C, attributable to loss of lattice water and an enthalpy of transition of 328.7 ± 5.7 J/g (mean ± standard deviation; n = 3) in the DSC, confirmed that the sample consisted of ∼60% DCPD. Dehydration Under Isothermal Conditions
At Different Water Vapor Pressures To systematically investigate the effect of externally applied water vapor pressure on dehydration, DCPD was subjected to a range of RH, 0–80%, at a fixed temperature of 60◦ C. From the measured weight loss, the percent dehydration was calculated. This was DOI 10.1002/jps
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
1459
Figure 1. Representative differential scanning calorimeter and thermogravimetric analyzer (TGA) plots of dibasic calcium phosphate dihydrate heated in (a) an open pan and (b) a hermetic pan with a pinhole. The derivative weight (derivative TGA plot) as a function of temperature is also shown.
performed assuming that the processes of dehydration and removal (vaporization) of the liberated water were occurring concomitantly. In light of the high nitrogen flow rate (75 mL/min), this is a reasonable assumption. Between 40% and 80% RH, the dehyDOI 10.1002/jps
dration was rapid and complete in less than 1 day (Fig. 3a). At RH values <40%, there was a progressive reduction in dehydration rate with decrease in RH. However, the difference between the rate of dehydration at 10% RH and under a dry nitrogen purge JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1460
KAUSHAL, VANGALA, AND SURYANARAYANAN
Figure 2. X-ray diffractometer patterns obtained when dibasic calcium phosphate dihydrate (DCPD) was subjected to a controlled temperature program (a) in a sealed capillary and (b) under reduced pressure (50 mTorr). Numbers next to diffraction patterns indicate the temperature in ◦ C. One characteristic peak, each of DCPD and dibasic calcium phosphate anhydrate (DCPA), is indicated. In sealed capillary, DCPA was readily discernible at 80◦ C and the peaks attributable to DCPD had disappeared at 100◦ C. Under reduced pressure, although reduced in intensity, the DCPD peaks were present even at 300◦ C.
(0% RH) was not very pronounced. Because the dehydration profiles were nonlinear, we plotted the time taken for 10% and 50% dehydration (weight loss) as a function of RH (Fig. 3b). This permits convenient, although not a comprehensive visualization of dehydration kinetics. At 60◦ C, it is evident that dehydration occurs most rapidly at RH ≥ 40%.
Under “Dry” Conditions As is evident from Figure 3a, at 60◦ C, dehydration occurred very slowly under dry nitrogen purge (0% RH). To determine the effect of temperature, DCPD JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
was subjected isothermally to a range of temperatures from 60◦ C to 250◦ C under a dry nitrogen purge. At temperatures ≥190◦ C, dehydration was very rapid and the weight loss was complete in ∼5 min (Fig. 4a). As the temperature of isothermal dehydration was lowered, there was a progressive decrease in dehydration rate (Figs. 4a and 4b). At 150◦ C, the initial weight loss (corresponding to ∼25% dehydration) was very rapid (<5 min), followed by a pronounced decrease in dehydration rate. Thus, after 15 h, there was ∼45% dehydration (Fig. 4b); and even after 6 days, only ∼60% of DCPD had dehydrated (not shown). The DOI 10.1002/jps
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
1461
Figure 3. (a) Isothermal (60◦ C) thermogravimetric analyzer plots of dibasic calcium phosphate dihydrate at different relative humidity (RH) values. (b) Time taken for 10% (-•-) and 50% (--) dehydration as a function of RH.
resistance to complete dehydration even at a high temperature of 150◦ C was surprising and was in sharp contrast to the rapid and complete dehydration at 60◦ C in the presence of water vapor.
Isothermal XRD in a Sealed Capillary The purpose of this experiment was to evaluate the influence of a “sealed” environment on the dehydration behavior of DCPD. Additionally, this would simuDOI 10.1002/jps
late water impermeable packaging of a tablet dosage form. In this configuration, any water released by the solid would dictate the water vapor pressure in the headspace. When held at 60◦ C, there was a progressive decrease in the intensity of the DCPD peaks and they completely disappeared after 5 days. At 40◦ C, although dehydration was discernible from the characteristic peaks of DCPA, the process was very slow. Even after 30 days, <20% of the DCPD had dehydrated. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1462
KAUSHAL, VANGALA, AND SURYANARAYANAN
Figure 4. Isothermal thermogravimetric analyzer plots of dibasic calcium phosphate dihydrate (in absence of water vapor) over different temperature ranges: (a) 150– 250◦ C and (b) 60–150◦ C.
Unusual Dehydration Behavior of DCPD The dehydration behavior of DCPD reveals several interesting and unexpected results: (i) The type of DSC pan used had a profound influence on the dehydration temperature and kinetics. Dehydration was facilitated when DCPD was heated in a hermetically sealed pan with a JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
pinhole, wherein the water removed by dehydration can be in intimate contact with the product (Figs. 1a and 1b, Table 1). (ii) Variable-temperature XRD aided in the confirmation of DSC results. When heated in a sealed capillary, dehydration was complete by 100◦ C, whereas heating in an “open holder” (the water vapor pressure above the sample is expected to be very low) caused complete dehydration only DOI 10.1002/jps
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
Table 1.
1463
Effect of Experimental Conditions on Dehydration Onset Temperature of DCPD Temperatures of Onset (and Completion) of Dehydration (◦ C) Variable Temperature XRDa
DSC and TGA Sample
Open Pan
Hermetic Panb
Ambient Pressure
Reduced Pressure
Sealed Capillary
“As is” DCPD Partially dehydrated DCPDd Partially dehydrated DCPDe Partially dehydrated DCPDd with added water
85 (210) 165 (210) 165 (210) 165 (210)
85 (140) 165 (210) 165 (210) 85 (140)
100 (200) – – –
200 (300)c – – –
80 (100) – – –
The temperature of completion of dehydration is given in parenthesis. a Dehydration onset temperature based on the discernibility of 26.4◦ 22 peak of DCPA. b Hermetic pan with a 70 :m pin hole. c Dehydration was not complete even at 300◦ C. d Prepared by heating DCPD in a variable-temperature XRD to 300◦ C under reduced pressure. e Prepared by heating DCPD (in a TGA) at 150◦ C for 12 h. DCPA, dibasic calcium phosphate anhydrate; DCPD, dibasic calcium phosphate dihydrate; DSC, differential scanning calorimeter; TGA, thermogravimetric analyzer; XRD, X-ray diffractometer.
at 220◦ C. Finally, when heated under reduced pressure (wherein the liberated water will be rapidly removed), dehydration was incomplete even at 300◦ C (Figs. 2a and 2b, Table 1). (iii) At 60◦ C, when exposed to RH values ranging from 0% to 80%, dehydration occurred most rapidly at RH values ≥40% (Fig. 3a) and was complete in <1 day. (iv) On the contrary, in the absence of water vapor (nitrogen purge), when held at 150◦ C, dehydration was substantially incomplete (∼60% dehydration) even after 6 days (Fig. 4b). Smith–Topley Effect In a number of inorganic hydrates, including manganese oxalate dihydrate,23,24 lithium sulfate monohydrate,25 and copper sulfate pentahydrate,26 over a limited range of water vapor pressure, an increase in water vapor pressure promoted dehydration. This has been referred to as the Smith–Topley effect. According to the elegant mechanistic interpretation provided by Galwey and Brown,27 dehydration is believed to result in a poorly crystalline product that can impede further dehydration. However, in the presence of water vapor, crystallization of the product is facilitated. This is understandable in light of the fact that water is a plasticizer, and its sorption can lead to crystallization of the product layer. Crystallization, through the formation of cracks and opening up of channels, causes a change in the nature of the product layer and facilitates water removal. We postulate that under conditions of low water vapor pressure, the poorly crystalline anhydrate formed by dehydration of DCPD “coats” the particle surface and impedes escape of water, thereby preventing further dehydration. On the contrary, when the water vapor pressure is high, the poorly crystalline DCPA rapidly crystallizes, leading to the formation of cracks and the opening of channels facilitating water reDOI 10.1002/jps
moval and continuation of dehydration. Although we have no direct evidence for the formation of a poorly crystalline product layer on the surface of dehydrating DCPD particles, there is substantial circumstantial evidence to support this postulate. We will now attempt to explain the experimental results, assuming the formation of a disordered layer on the surface of dehydrating particles.
Effect of Water Vapor Pressure on Dehydration Kinetics As pointed out earlier, dehydration was facilitated in hermetically sealed pans with pinhole (Fig. 1b). An enlarged view of the DSC heat flow curves shows that, in both open and sealed pan configurations, the dehydration started at approximately the same temperature of ∼85◦ C (Fig. 5). However, in open pans, the water released upon dehydration is removed readily, retaining the poorly crystalline DCPA surface layer. This layer inhibited, though it did not completely prevent, further dehydration until ∼175◦ C. Dehydration was very rapid at temperatures >175◦ C. In the hermetic pan configuration, the intimate contact of the water released upon dehydration with the product will facilitate crystallization of the disordered surface layer, leading to the formation of cracks and channels and allowing the dehydration to proceed to completion at a much lower temperature of ∼140◦ C. The dehydration behavior of DCPD in a variable-temperature XRD, using either a sealed capillary (Fig. 2a) or in an open holder (data not shown), can be similarly explained. Additional supporting evidence was obtained when the DCPD was heated under reduced pressure, so as to rapidly remove the liberated water. By inhibiting the crystallization of the disordered surface layer, dehydration was substantially incomplete even up to 300◦ C (Fig. 2b). Both in the DSC (hermetic pan with pinhole) and XRD (sealed capillary) experiments, following the initial dehydration, the solid will experience an JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1464
KAUSHAL, VANGALA, AND SURYANARAYANAN
Figure 5. Enlarged view of differential scanning calorimeter heating curve of dibasic calcium phosphate dihydrate in an open pan and a hermetic pan with pinhole.
elevated but uncontrolled water vapor pressure. On the contrary, the automated water sorption apparatus permitted a systematic evaluation of the effect of water vapor pressure on dehydration. The resistance to dehydration, following exposure to low RH values (<20%, 60◦ C), can be attributed to the physical stability of the surface disordered layer (Fig. 3a). This surface layer, once formed, appears to be effective in inhibiting dehydration even at elevated temperatures. For example, when DCPD was exposed to 150◦ C, the initial rapid weight loss can be attributed to the formation of the poorly crystalline surface DCPA layer. This layer then effectively inhibited further dehydration for a prolonged time period. Thus, even after 6 days, only ∼60% of DCPD had dehydrated (data not shown).
hydrated at ∼210◦ C as opposed to ∼140◦ C for the “as is” sample (Table 1), lending credence to the formation of a poorly crystalline product layer on the surface of particles during heating in the XRD (absence of water vapor in the atmosphere). Once the poorly crystalline layer was formed, the type of DSC pan had no effect on the temperature of dehydration. In contrast, the dehydration behavior of “as is” DCPD was strongly influenced by the type of pan used (Figs. 1a and 1b). To confirm these results, DCPD was also partially dehydrated (∼45% dehydration) by holding isothermally at 150◦ C for 12 h and subjected to DSC in a hermetically crimped pan with a pinhole. Again, the dehydration was complete by 210◦ C (Table 1).
“Recrystallization” of Poorly Crystalline Surface Layer Poorly Crystalline Surface Layer—Effect on Dehydration Thus, we have strong circumstantial evidence for the formation of a disordered layer on the surface of dehydrating particles. We wanted to directly examine the effect of the disordered surface layer on the dehydration behavior. This can only be performed in DCPD samples, which have been partially dehydrated, in the absence of water vapor. As described earlier, DCPD dehydrated partially (∼40% DCPA) when heated to 300◦ C, in a variable-temperature XRD, under reduced pressure (Fig. 2b). This sample was cooled back to room temperature and subjected to DSC, both in an open and in a hermetically sealed pan with a pinhole. In an open pan, the temperature of complete dehydration for the partially dehydrated sample was similar (∼210◦ C) to the “as is” DCPD. However, in a hermetic pan with pinhole, the partially dehydrated sample deJOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
Finally, we attempted to recrystallize the disordered surface layer by adding water to partially dehydrated DCPD. In this experiment, ∼1.5 mg of water was added to ∼7 mg of DCPD that had been previously dehydrated partially by holding it at 150◦ C for 12 h. This sample, when subjected to DSC in a hermetic pan with pinhole, exhibited an endotherm starting at ∼85◦ C, attributable to dehydration (Table 1, figure not shown). Thus, the ability of the poorly crystalline surface layer to inhibit dehydration could be effectively “reversed” by recrystallization of this layer. The dehydration studies carried out under a variety of conditions strongly suggest that, in the absence of water vapor, a poorly crystalline layer is formed on the surface of the dehydrating DCPD particles. This was confirmed by detailed thermal characterization of the partially dehydrated material. DOI 10.1002/jps
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
Pharmaceutical Significance DCPD is widely used as a diluent in tablets. As mentioned earlier, when the weight fraction of the API is low, excipients, especially diluents, constitute a major proportion of the dosage form. As DCPD has 20.9% (w/w) water, dehydration of even a small fraction can release a substantial amount of water. This is not unlikely in light of the fact that dehydration can occur under pharmaceutically relevant storage conditions. Unlike “classical” hydrates, an elevated water vapor pressure facilitated the dehydration of DCPD. Thus, if a DCPD containing formulation is in a well-packaged container, once dehydration is initiated, it can lead to buildup of water vapor, and the consequent increase in the water vapor pressure can accelerate further dehydration. This was apparent from the experiment in which complete dehydration was observed when DCPD was held isothermally at 60◦ C for 5 days in a sealed glass capillary. In solid dosage forms, if the water released by the dehydration of DCPD is sorbed by the other formulation components, there will be no change in the total water content. However, the released water, in light of its mobility, may bring about undesirable physical and chemical changes. This is the subject of a current investigation in our laboratory.
1465
REFERENCES
We gratefully acknowledge the financial support from Dane O. Kildsig Center for Pharmaceutical Processing Research. Parts of this work were carried out in the Characterization Facility, University of Minnesota, which receives partial support from the National Science Foundation through the Materials Research Science and Engineering Centers program.
1. Debnath S Suryanarayanan R. 2004. Influence of processinginduced phase transformations on the dissolution of theophylline tablets. AAPS PharmSciTech 5:39–49. 2. Govindarajan R Suryanarayanan R. 2006. Processing- induced phase transformations and their implications on pharmaceutical products quality. In Polymorphism; Hilfiker R, Ed., Weinheim, Germany: Wiley-VCH, pp 333–364. 3. Sakata Y, Shiraishi S, Takayama K, Otsuka M. 2006. Effect of pulverization and dehydration on the pharmaceutical properties of calcium lactate pentahydrate tablets. Colloids Surf B Biointerfaces 51:149–156. 4. Otsuka M, Kaneniwa N, Otsuka K, Kawakami K, Umezawa O. 1993. Effect of tableting pressure and geometrical factor of tablet on dehydration kinetics of theophylline monohydrate tablets. Drug Dev Ind Pharm 19:541–557. 5. Moreton R. 2000. Calcium phosphate dibasic dihydrate. In Handbook of pharmaceutical excipients; Kibbe AH, Ed. London, UK: American Pharmaceutical Association, pp 63–67. 6. Moreton R. 2000. Calcium phosphate dibasic anhydrous. In Handbook of pharmaceutical excipients; Kibbe AH, Ed. London, UK: American Pharmaceutical Association, pp 60–62. 7. Schmidt PC, Herzog R. 1993. Calcium phosphates in pharmaceutical tableting. 1. Physico-pharmaceutical properties. Pharm World Sci 15:105–115. 8. Schmidt PC, Herzog R. 1993. Calcium phosphates in pharmaceutical tableting. 2. Comparison of tableting properties. Pharm World Sci 15:116–122. 9. Gohel MC, Parikh RK, Brahmbhatt BK, Shah AR. 2007. Preparation and assessment of novel coprocessed superdisintegrant consisting of crospovidone and sodium starch glycolate. AAPS PharmSciTech 8:E63–E69. 10. Landin M, Casalderrey M, Martinez-Pacheco R, Gomez-Amoza JL, Souto C, Concheiro A, Rowe RC. 1995. Chemical stability of acetylsalicylic acid in tablets prepared with different particle size fractions of a commercial brand of dicalcium phosphate dihydrate. Int J Pharm 123:143–144. 11. Duff EJ. 1971. Orthophosphates. XIII. Thermal decomposition of secondary calcium orthophosphate (CaHPO4 ) and secondary calcium orthophosphate dihydrate (CaHPO4 •2H2 O). J Appl Chem Biotechnol 21:233–235. 12. Dugleux P, De Sallier Dupin A. 1967. Kinetic studies of the dehydration of dicalcium orthophosphate. III. Evaluation of datascheme of reaction mechanism. Bull Soc Chim Fr 3:978–982. 13. Dugleux P, De Sallier Dupin A. 1967. Kinetic studies of the dehydration of dicalcium orthophosphate. II. Dehydration of calcium hydrogen phosphate dihydrate in an anhydrous or slightly wet atmosphere. Bull Soc Chim Fr 3:973–977. 14. Landin M, Rowe RC, York P. 1994. Structural changes during the dehydration of dicalcium phosphate dihydrate. Eur J Pharm Sci 2:245–252. 15. Landin M, Fontao MJ, Martinez-Pacheco R. 2005. A comparison of trehalose dihydrate and mannitol as stabilizing agents for dicalcium phosphate dihydrate based tablets. Drug Dev Ind Pharm 31:249–256. 16. Horhota ST, Burgio J, Lonski L, Rhodes CT. 1976. Effect of storage at specified temperature and humidity on properties of three directly compressible tablet formulations. J Pharm Sci 65:1746–1749. 17. Lausier JM, Chiang C-W, Zompa HA, Rhodes CT. 1977. Aging of tablets made with dibasic calcium phosphate dihydrate as matrix. J Pharm Sci 66:1636–1637. 18. Miyazaki T, Sivaprakasam K, Tantry J, Suryanarayanan R. 2009. Physical characterization of dibasic calcium phosphate dihydrate and anhydrate. J Pharm Sci 98:905–916. 19. Han J, Suryanarayanan R. 1997. Applications of pressure differential scanning calorimetry in the study of
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
CONCLUSIONS Dibasic calcium phosphate dihydrate, a widely used tablet diluent, dehydrates over pharmaceutically relevant timescales and conditions. The dehydration behavior is unusual in that it is accelerated in the presence of water vapor. The increase in dehydration rate with the increase in water vapor pressure can be rationalized if we consider the formation of a poorly crystalline anhydrate layer on the surface of dehydrating particles of DCPD. “Dry” conditions favor existence of the poorly crystalline layer and impede dehydration. On the contrary, in the presence of water vapor, recrystallization occurs, creating cracks and channels and facilitating removal of water. In a well-packaged solid dosage form, any water released upon initial dehydration is likely to be “trapped” and can accelerate further dehydration with potential implications on product attributes and performance.
ACKNOWLEDGMENTS
1466
20. 21.
22. 23.
KAUSHAL, VANGALA, AND SURYANARAYANAN
pharmaceutical hydrates. I. Carbamazepine dihydrate. Int J Pharm 157:209–218. Suryanarayanan R, Mitchell AG. 1986. Phase transitions of calcium gluceptate. Int J Pharm 32:213–221. Willart JF, De Gusseme A, Hemon S, Descamps M, Leveiller F, Rameau A. 2002. Vitrification and polymorphism of trehalose induced by dehydration of trehalose dihydrate. J Phys Chem B 106:3365–3370. Zhu H, Grant DJW. 2001. Dehydration behavior of nedocromil magnesium pentahydrate. Int J Pharm 215:251–262. Smith ML, Topley B. 1932. Heats of formation of the hydrates of manganese oxalate. J Chem Soc 1977–1979.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
24. Topley B, Smith ML. 1935. Kinetics of salt-hydrate dissociations: MnC2 O4 .2H2 O = MnC2 O4 + 2H2 O. J Chem Soc 57:321–325. 25. Seto Y, Sato H, Masuda Y. 2002. Effect of water vapor pressure on thermal dehydration of lithium sulfate monohydrate. Thermochim Acta 388:21–25. 26. Bertrand G, Lallemant M, Watelle-Marion G. 1974. Abnormal variation of the rate of decomposition of a solid. I. Dehydration of salt hydrates. J Inorg Nucl Chem 36:1303–1309. 27. Galwey A, Brown M. 1999. Thermal decomposition of ionic solids. Amsterdam, the Netherlands: Elsevier Science BV, pp 217–268.
DOI 10.1002/jps
Department of Pharmaceutics, College of Pharmacy, University of Minnesota, Minneapolis, Minnesota 55455
2 Venu R. Vangala’s present address is Institute of Chemical and Engineering Sciences, A∗STAR (Agency for Science, Technology and Research), Jurong Island, Singapore 627833
Received 17 August 2010; revised 8 September 2010; accepted 20 September 2010 Published online 6 December 2010 in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jps.22372 ABSTRACT: Dibasic calcium phosphate occurs as an anhydrate (DCPA; CaHPO4 ) and as a dihydrate (DCPD; CaHPO4 •2H2 O). Our objective was to investigate the unusual behavior of these phases. Dibasic calcium phosphate dihydrate was dehydrated in a (i) differential scanning calorimeter (DSC) in different pan configurations; (ii) variable-temperature X-ray diffractometer (XRD) at atmospheric and under reduced pressure, and in sealed capillaries; and (iii) water vapor sorption analyzer at varying temperature and humidity conditions. Dehydration was complete by 210◦ C in an open DSC pan and under atmospheric pressure in the XRD. Unlike “conventional” hydrates, the dehydration of DCPD was facilitated in the presence of water vapor. Variable-temperature XRD in a sealed capillary and DSC in a hermetic pan with pinhole caused complete dehydration by 100◦ C and 140◦ C, respectively. Under reduced pressure, conversion to the anhydrate was incomplete even at 300◦ C. The increase in dehydration rate with increase in water vapor pressure has been explained by the Smith–Topley effect. Under “dry” conditions, a coating of poorly crystalline product is believed to form on the surface of particles and act as a barrier to further dehydration. However, in the presence of water vapor, recrystallization occurs, creating cracks and channels and facilitating continued dehydration. © 2010 Wiley-Liss, Inc. and the American Pharmacists Association J Pharm Sci 100:1456–1466, 2011 Keywords: dibasic calcium phosphate; dehydration; water vapor; calorimetry (DSC); thermogravimetric analysis; X-ray diffractometry; excipients; hydrate; physical stability
INTRODUCTION Solid oral dosage forms typically consist of an active pharmaceutical ingredient (API) and multiple excipients. The performance of the dosage form may be influenced by the physical forms of the formulation components. A large number of studies have focused on the physical form [the term “physical form” encompasses polymorphs, solvates, the noncrystalline (amorphous) form, and partially crystalline forms] of APIs, both during the manufacture of the dosage form and its subsequent storage. The implications of such phase transformations, both during pharmaceutical processing and subsequent storage, on product performance have also been evaluated.1,2 For example, a number of APIs are marketed as hydrates, wherein Correspondence to: Raj Suryanarayanan (Tel: 612-624-9626; Fax: 612-626-2125; E-mail: [email protected]) Journal of Pharmaceutical Sciences, Vol. 100, 1456–1466 (2011) © 2010 Wiley-Liss, Inc. and the American Pharmacists Association
1456
water molecules are incorporated, usually stoichiometrically, in the crystal lattice. Dehydration during storage can have implications on product performance and this issue has received considerable attention.3,4 However, such in-depth attention has not been paid to the phase behavior of excipients. When the weight fraction of the API in the dosage form is low, excipients, especially diluents, constitute a major proportion of the dosage form. Phase transformations of excipients either during processing or during subsequent storage can be equally serious because they can also influence the dosage form performance. Dibasic calcium phosphate is widely used as a diluent in tablets and capsules, and occurs as dibasic calcium phosphate anhydrate (DCPA; CaHPO4 ) and as dibasic calcium phosphate dihydrate (DCPD; CaHPO4 •2H2 O).5–8 Both DCPD and DCPA exhibit good compressibility, amenability to scale-up, and brittle fracture propensity, making them popular excipients. Dibasic calcium phosphate dihydrate has
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1457
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
a stoichiometric water content of 20.9% (w/w). As a diluent, it can constitute a substantial fraction (over 50%, w/w) of the dosage form.9,10 The dehydration of DCPD to yield DCPA has been documented in numerous studies.11–14 Given its high water content, dehydration of even a fraction of DCPD in the formulation can lead to the release of an appreciable amount of water. The released water can cause chemical or physical changes in the dosage form components. Aspirin degradation during accelerated stability testing [35◦ C and 82.9% relative humidity (RH)] has been attributed to the water released by DCPD dehydration.10,15 In addition, the functionality of excipients can be affected by the released water. For example, in superdisintegrants, the sorption of the released water can cause product failure. Finally, changes in tablet hardness and dimensions were attributed to in situ DCPD dehydration.16,17 It is important to recognize that DCPD dehydration in situ may not be readily apparent, if the liberated water is taken up by one or more formulation components. There will not be any change in tablet weight because water is transferred from one formulation component to another. In other words, the effect is brought about by a change in the distribution of water in the dosage form. The dehydration behavior of DCPD alone18 was evaluated as the first step toward understanding the dehydration in complex multicomponent formulations. When heated in a thermogravimetric analyzer (TGA) under nitrogen purge, the dehydration was complete by 210◦ C. When DCPD was dehydrated at 40◦ C, but at RH values ranging from 2% to 75%, the dehydration kinetics was heavily influenced by the water vapor pressure. The most rapid dehydration, based on the observed weight loss, occurred at intermediate RH values of 32% and 52% RH. On the contrary, DCPA was remarkably resistant to hydration, and it was stable even when dispersed in water at 50◦ C for 7 months.18 A conventional anhydrate–hydrate pair, under isothermal conditions, can exist in stable equilibrium at the transition water vapor pressure (pt ), and the anhydrate is stable at water vapor pressures
in a TGA, differential scanning calorimeter (DSC), and in an X-ray diffractometer (XRD), it was possible to comprehensively characterize the reactant (DCPD) and the product (DCPA) phases.
EXPERIMENTAL Materials R
DCPD from a commercial source (Emcompress batch 7070X, JRS Pharma, Patterson, New York) was used for the work. Differential Scanning Calorimetry The sample (5–10 mg) was heated in a DSC (Model 2920, TA Instruments, New Castle, Delaware) from 25◦ C to 250◦ C at 10◦ C/min under nitrogen purge (75 mL/min). Two different pan configurations, open pan (no lid) and hermetically sealed pan with standard 75 :m pinhole, were used. The experimental data were analyzed using a commercially available software (TA Universal Analysis 2000; TA Instruments). Thermogravimetric Analysis TGA was employed in both isothermal and nonisothermal modes. For the nonisothermal analysis, 5–10 mg of the sample was heated in the TGA (Model Q50, TA Instruments) from room temperature to 250◦ C at 10◦ C/min under nitrogen purge (75 mL/min). Both the pan configurations, as in the DSC, were utilized. For the isothermal analyses, the temperature was rapidly increased to the desired value and maintained isothermally thereafter. Variable-Temperature X-Ray Diffractometry
Under Reduced Pressure Approximately 100 mg of sample was filled in an aluminum holder by the top-fill technique and exposed to CuK" radiation (45 kV × 40 mA) in a wide-angle XRD (model XDS 2000, Scintag Inc., Cupertino, California). The sample holder was covered with a stainless steel dome having a beryllium window and evacuated R to ∼50 mTorr. A temperature controller (Micristar , Model 828D; TMC Services, Inc., Elk River, Minnesota) was used to heat the sample from 25◦ C to 300◦ C. The angular range was 5◦ to 40◦ 22 and counts were accumulated for 1 s at each step. During the XRD run (∼12 min), the sample was maintained under isothermal conditions. Data analyses were performed using commercially available software (JADE; Materials Data, Inc., Livermore, California).
In Sealed Capillaries DCPD was sealed in a glass capillary (0.7 mm diameter; Charles Supper Company, Natick, Massachusetts) and exposed to CuK" radiation JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1458
KAUSHAL, VANGALA, AND SURYANARAYANAN
(50 kV × 100 mA; collimated to 0.5 mm i.d. spot size) generated with a rotating anode (18 kW; Rotaflex generator, Woodlands, Texas). A two-dimensional area detector (Highstar, D8 Discover Platform; Bruker AXS, Madison, Wisconsin) was used at a fixed position of 22◦ 22, which covered the angular range from 7◦ to 37◦ 22. Capillaries were heated from 25◦ C to 100◦ C using a temperature controller (3000CN, Omega Engineering, Stamford, Connecticut).
tion of hydrates.19,22 However, the opposite effect was observed here indicating that dehydration of DCPD was facilitated in the presence of water vapor. The enthalpy of transition (dehydration accompanied by water vaporization) was nearly identical in the two pan configurations: 567.5 ± 5.4 (mean ± standard deviation; n = 4) in open pans and 570.9 ± 4.7 (mean ± standard deviation; n = 4) in hermetic pans with pinhole.
Dehydration in an Automated Water Sorption/Desorption Apparatus
Variable-Temperature XRD
Approximately 20 mg of the sample was placed in the quartz sample pan of an automated vapor sorption analyzer (TGA Q5000SA; TA Instruments) and equilibrated at 60◦ C/0% RH. The RH was then rapidly changed to a fixed value (ranging from 0% to 80%, 60◦ C), where the sample was held for 1800 min and the weight was continuously monitored.
RESULTS AND DISCUSSION The dehydration behavior of DCPD was evaluated under (i) nonisothermal and (ii) isothermal conditions. Dehydration Under Nonisothermal Conditions
Differential Scanning Calorimetry and Thermogravimetric Analysis The DSC curve of DCPD, when heated in an open pan, was substantially similar to the profile reported earlier.18 The three endotherms were attributed to dehydration accompanied by vaporization of water (Fig. 1a). This conclusion was based on the fact that in the temperature range where the endotherms were observed in the DSC, a weight loss was observed in the TGA. The derivative TGA plots enabled us to discern the precise temperature range of the weight loss. Dehydration appeared to be complete by ∼210◦ C. The thermal behavior of DCPD, characterized by DSC, TGA, and variable-temperature XRD, was discussed in our earlier publication.18 In our previous DSC studies, only open pans (without lids) were used. It is well known that the dehydration behavior of hydrates can be influenced by the DSC pan configuration.19–21 In a hermetically sealed pan with a pinhole, the released water will not leave the pan until the water vapor pressure in the pan is substantially greater than the atmospheric pressure. Therefore, the solid will experience an elevated, though uncontrolled water vapor pressure. In this configuration, the dehydration of DCPD appeared to be a single-step process and was complete by ∼140◦ C (Fig. 1b). The weight loss observed in a TGA, also in a hermetically sealed pan with a pinhole, enabled us to confirm these results. Conventionally, an increase in water vapor pressure will decelerate the dehydraJOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
In an effort to obtain phase information, DCPD was subjected to variable-temperature XRD under ambient pressure. As reported previously, DCPD directly and completely dehydrated to DCPA by ∼220◦ C with no evidence of any intermediate phases.18 Even at a rapid heating rate of 40◦ C/min, no halos, attributable to amorphous phases, were seen. Because we could not simulate the exact DSC pan configuration (sealed pan with a pinhole), DCPD was subjected to variabletemperature XRD in a sealed capillary. When heated in a sealed capillary, dehydration appeared to be complete by ∼100◦ C (disappearance of the 21.0◦ 22peak of DCPD) with the concomitant appearance of the DCPA peaks at 26.4◦ and 30.2◦ 22 (Fig. 2b). In this configuration, the water released by dehydration can be in intimate contact with the solid and this appeared to substantially accelerate the dehydration of DCPD. In an effort to confirm the role of water vapor, DCPD was heated under reduced pressure (∼50 mTorr). Under this condition, the water vapor released upon dehydration would be quickly removed, thereby limiting the interaction between water vapor and solid. Interestingly, the dehydration of DCPD was incomplete even at 300◦ C (Fig. 2b). Although the characteristic peak of DCPA at 26.4◦ 22 increased in intensity with temperature, the DCPD peaks (21.0◦ and 11.7◦◦ 22) remained detectable through the duration of the experiment. To verify that the dehydration was incomplete at 300◦ C, the solid was cooled to room temperature and subjected to TGA and DSC in an open pan. In the TGA, a sharp weight loss of 11.6% at ∼175◦ C, attributable to loss of lattice water and an enthalpy of transition of 328.7 ± 5.7 J/g (mean ± standard deviation; n = 3) in the DSC, confirmed that the sample consisted of ∼60% DCPD. Dehydration Under Isothermal Conditions
At Different Water Vapor Pressures To systematically investigate the effect of externally applied water vapor pressure on dehydration, DCPD was subjected to a range of RH, 0–80%, at a fixed temperature of 60◦ C. From the measured weight loss, the percent dehydration was calculated. This was DOI 10.1002/jps
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
1459
Figure 1. Representative differential scanning calorimeter and thermogravimetric analyzer (TGA) plots of dibasic calcium phosphate dihydrate heated in (a) an open pan and (b) a hermetic pan with a pinhole. The derivative weight (derivative TGA plot) as a function of temperature is also shown.
performed assuming that the processes of dehydration and removal (vaporization) of the liberated water were occurring concomitantly. In light of the high nitrogen flow rate (75 mL/min), this is a reasonable assumption. Between 40% and 80% RH, the dehyDOI 10.1002/jps
dration was rapid and complete in less than 1 day (Fig. 3a). At RH values <40%, there was a progressive reduction in dehydration rate with decrease in RH. However, the difference between the rate of dehydration at 10% RH and under a dry nitrogen purge JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1460
KAUSHAL, VANGALA, AND SURYANARAYANAN
Figure 2. X-ray diffractometer patterns obtained when dibasic calcium phosphate dihydrate (DCPD) was subjected to a controlled temperature program (a) in a sealed capillary and (b) under reduced pressure (50 mTorr). Numbers next to diffraction patterns indicate the temperature in ◦ C. One characteristic peak, each of DCPD and dibasic calcium phosphate anhydrate (DCPA), is indicated. In sealed capillary, DCPA was readily discernible at 80◦ C and the peaks attributable to DCPD had disappeared at 100◦ C. Under reduced pressure, although reduced in intensity, the DCPD peaks were present even at 300◦ C.
(0% RH) was not very pronounced. Because the dehydration profiles were nonlinear, we plotted the time taken for 10% and 50% dehydration (weight loss) as a function of RH (Fig. 3b). This permits convenient, although not a comprehensive visualization of dehydration kinetics. At 60◦ C, it is evident that dehydration occurs most rapidly at RH ≥ 40%.
Under “Dry” Conditions As is evident from Figure 3a, at 60◦ C, dehydration occurred very slowly under dry nitrogen purge (0% RH). To determine the effect of temperature, DCPD JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
was subjected isothermally to a range of temperatures from 60◦ C to 250◦ C under a dry nitrogen purge. At temperatures ≥190◦ C, dehydration was very rapid and the weight loss was complete in ∼5 min (Fig. 4a). As the temperature of isothermal dehydration was lowered, there was a progressive decrease in dehydration rate (Figs. 4a and 4b). At 150◦ C, the initial weight loss (corresponding to ∼25% dehydration) was very rapid (<5 min), followed by a pronounced decrease in dehydration rate. Thus, after 15 h, there was ∼45% dehydration (Fig. 4b); and even after 6 days, only ∼60% of DCPD had dehydrated (not shown). The DOI 10.1002/jps
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
1461
Figure 3. (a) Isothermal (60◦ C) thermogravimetric analyzer plots of dibasic calcium phosphate dihydrate at different relative humidity (RH) values. (b) Time taken for 10% (-•-) and 50% (--) dehydration as a function of RH.
resistance to complete dehydration even at a high temperature of 150◦ C was surprising and was in sharp contrast to the rapid and complete dehydration at 60◦ C in the presence of water vapor.
Isothermal XRD in a Sealed Capillary The purpose of this experiment was to evaluate the influence of a “sealed” environment on the dehydration behavior of DCPD. Additionally, this would simuDOI 10.1002/jps
late water impermeable packaging of a tablet dosage form. In this configuration, any water released by the solid would dictate the water vapor pressure in the headspace. When held at 60◦ C, there was a progressive decrease in the intensity of the DCPD peaks and they completely disappeared after 5 days. At 40◦ C, although dehydration was discernible from the characteristic peaks of DCPA, the process was very slow. Even after 30 days, <20% of the DCPD had dehydrated. JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1462
KAUSHAL, VANGALA, AND SURYANARAYANAN
Figure 4. Isothermal thermogravimetric analyzer plots of dibasic calcium phosphate dihydrate (in absence of water vapor) over different temperature ranges: (a) 150– 250◦ C and (b) 60–150◦ C.
Unusual Dehydration Behavior of DCPD The dehydration behavior of DCPD reveals several interesting and unexpected results: (i) The type of DSC pan used had a profound influence on the dehydration temperature and kinetics. Dehydration was facilitated when DCPD was heated in a hermetically sealed pan with a JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
pinhole, wherein the water removed by dehydration can be in intimate contact with the product (Figs. 1a and 1b, Table 1). (ii) Variable-temperature XRD aided in the confirmation of DSC results. When heated in a sealed capillary, dehydration was complete by 100◦ C, whereas heating in an “open holder” (the water vapor pressure above the sample is expected to be very low) caused complete dehydration only DOI 10.1002/jps
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
Table 1.
1463
Effect of Experimental Conditions on Dehydration Onset Temperature of DCPD Temperatures of Onset (and Completion) of Dehydration (◦ C) Variable Temperature XRDa
DSC and TGA Sample
Open Pan
Hermetic Panb
Ambient Pressure
Reduced Pressure
Sealed Capillary
“As is” DCPD Partially dehydrated DCPDd Partially dehydrated DCPDe Partially dehydrated DCPDd with added water
85 (210) 165 (210) 165 (210) 165 (210)
85 (140) 165 (210) 165 (210) 85 (140)
100 (200) – – –
200 (300)c – – –
80 (100) – – –
The temperature of completion of dehydration is given in parenthesis. a Dehydration onset temperature based on the discernibility of 26.4◦ 22 peak of DCPA. b Hermetic pan with a 70 :m pin hole. c Dehydration was not complete even at 300◦ C. d Prepared by heating DCPD in a variable-temperature XRD to 300◦ C under reduced pressure. e Prepared by heating DCPD (in a TGA) at 150◦ C for 12 h. DCPA, dibasic calcium phosphate anhydrate; DCPD, dibasic calcium phosphate dihydrate; DSC, differential scanning calorimeter; TGA, thermogravimetric analyzer; XRD, X-ray diffractometer.
at 220◦ C. Finally, when heated under reduced pressure (wherein the liberated water will be rapidly removed), dehydration was incomplete even at 300◦ C (Figs. 2a and 2b, Table 1). (iii) At 60◦ C, when exposed to RH values ranging from 0% to 80%, dehydration occurred most rapidly at RH values ≥40% (Fig. 3a) and was complete in <1 day. (iv) On the contrary, in the absence of water vapor (nitrogen purge), when held at 150◦ C, dehydration was substantially incomplete (∼60% dehydration) even after 6 days (Fig. 4b). Smith–Topley Effect In a number of inorganic hydrates, including manganese oxalate dihydrate,23,24 lithium sulfate monohydrate,25 and copper sulfate pentahydrate,26 over a limited range of water vapor pressure, an increase in water vapor pressure promoted dehydration. This has been referred to as the Smith–Topley effect. According to the elegant mechanistic interpretation provided by Galwey and Brown,27 dehydration is believed to result in a poorly crystalline product that can impede further dehydration. However, in the presence of water vapor, crystallization of the product is facilitated. This is understandable in light of the fact that water is a plasticizer, and its sorption can lead to crystallization of the product layer. Crystallization, through the formation of cracks and opening up of channels, causes a change in the nature of the product layer and facilitates water removal. We postulate that under conditions of low water vapor pressure, the poorly crystalline anhydrate formed by dehydration of DCPD “coats” the particle surface and impedes escape of water, thereby preventing further dehydration. On the contrary, when the water vapor pressure is high, the poorly crystalline DCPA rapidly crystallizes, leading to the formation of cracks and the opening of channels facilitating water reDOI 10.1002/jps
moval and continuation of dehydration. Although we have no direct evidence for the formation of a poorly crystalline product layer on the surface of dehydrating DCPD particles, there is substantial circumstantial evidence to support this postulate. We will now attempt to explain the experimental results, assuming the formation of a disordered layer on the surface of dehydrating particles.
Effect of Water Vapor Pressure on Dehydration Kinetics As pointed out earlier, dehydration was facilitated in hermetically sealed pans with pinhole (Fig. 1b). An enlarged view of the DSC heat flow curves shows that, in both open and sealed pan configurations, the dehydration started at approximately the same temperature of ∼85◦ C (Fig. 5). However, in open pans, the water released upon dehydration is removed readily, retaining the poorly crystalline DCPA surface layer. This layer inhibited, though it did not completely prevent, further dehydration until ∼175◦ C. Dehydration was very rapid at temperatures >175◦ C. In the hermetic pan configuration, the intimate contact of the water released upon dehydration with the product will facilitate crystallization of the disordered surface layer, leading to the formation of cracks and channels and allowing the dehydration to proceed to completion at a much lower temperature of ∼140◦ C. The dehydration behavior of DCPD in a variable-temperature XRD, using either a sealed capillary (Fig. 2a) or in an open holder (data not shown), can be similarly explained. Additional supporting evidence was obtained when the DCPD was heated under reduced pressure, so as to rapidly remove the liberated water. By inhibiting the crystallization of the disordered surface layer, dehydration was substantially incomplete even up to 300◦ C (Fig. 2b). Both in the DSC (hermetic pan with pinhole) and XRD (sealed capillary) experiments, following the initial dehydration, the solid will experience an JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
1464
KAUSHAL, VANGALA, AND SURYANARAYANAN
Figure 5. Enlarged view of differential scanning calorimeter heating curve of dibasic calcium phosphate dihydrate in an open pan and a hermetic pan with pinhole.
elevated but uncontrolled water vapor pressure. On the contrary, the automated water sorption apparatus permitted a systematic evaluation of the effect of water vapor pressure on dehydration. The resistance to dehydration, following exposure to low RH values (<20%, 60◦ C), can be attributed to the physical stability of the surface disordered layer (Fig. 3a). This surface layer, once formed, appears to be effective in inhibiting dehydration even at elevated temperatures. For example, when DCPD was exposed to 150◦ C, the initial rapid weight loss can be attributed to the formation of the poorly crystalline surface DCPA layer. This layer then effectively inhibited further dehydration for a prolonged time period. Thus, even after 6 days, only ∼60% of DCPD had dehydrated (data not shown).
hydrated at ∼210◦ C as opposed to ∼140◦ C for the “as is” sample (Table 1), lending credence to the formation of a poorly crystalline product layer on the surface of particles during heating in the XRD (absence of water vapor in the atmosphere). Once the poorly crystalline layer was formed, the type of DSC pan had no effect on the temperature of dehydration. In contrast, the dehydration behavior of “as is” DCPD was strongly influenced by the type of pan used (Figs. 1a and 1b). To confirm these results, DCPD was also partially dehydrated (∼45% dehydration) by holding isothermally at 150◦ C for 12 h and subjected to DSC in a hermetically crimped pan with a pinhole. Again, the dehydration was complete by 210◦ C (Table 1).
“Recrystallization” of Poorly Crystalline Surface Layer Poorly Crystalline Surface Layer—Effect on Dehydration Thus, we have strong circumstantial evidence for the formation of a disordered layer on the surface of dehydrating particles. We wanted to directly examine the effect of the disordered surface layer on the dehydration behavior. This can only be performed in DCPD samples, which have been partially dehydrated, in the absence of water vapor. As described earlier, DCPD dehydrated partially (∼40% DCPA) when heated to 300◦ C, in a variable-temperature XRD, under reduced pressure (Fig. 2b). This sample was cooled back to room temperature and subjected to DSC, both in an open and in a hermetically sealed pan with a pinhole. In an open pan, the temperature of complete dehydration for the partially dehydrated sample was similar (∼210◦ C) to the “as is” DCPD. However, in a hermetic pan with pinhole, the partially dehydrated sample deJOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
Finally, we attempted to recrystallize the disordered surface layer by adding water to partially dehydrated DCPD. In this experiment, ∼1.5 mg of water was added to ∼7 mg of DCPD that had been previously dehydrated partially by holding it at 150◦ C for 12 h. This sample, when subjected to DSC in a hermetic pan with pinhole, exhibited an endotherm starting at ∼85◦ C, attributable to dehydration (Table 1, figure not shown). Thus, the ability of the poorly crystalline surface layer to inhibit dehydration could be effectively “reversed” by recrystallization of this layer. The dehydration studies carried out under a variety of conditions strongly suggest that, in the absence of water vapor, a poorly crystalline layer is formed on the surface of the dehydrating DCPD particles. This was confirmed by detailed thermal characterization of the partially dehydrated material. DOI 10.1002/jps
EFFECT OF WATER VAPOR PRESSURE ON DEHYDRATION OF DCPD
Pharmaceutical Significance DCPD is widely used as a diluent in tablets. As mentioned earlier, when the weight fraction of the API is low, excipients, especially diluents, constitute a major proportion of the dosage form. As DCPD has 20.9% (w/w) water, dehydration of even a small fraction can release a substantial amount of water. This is not unlikely in light of the fact that dehydration can occur under pharmaceutically relevant storage conditions. Unlike “classical” hydrates, an elevated water vapor pressure facilitated the dehydration of DCPD. Thus, if a DCPD containing formulation is in a well-packaged container, once dehydration is initiated, it can lead to buildup of water vapor, and the consequent increase in the water vapor pressure can accelerate further dehydration. This was apparent from the experiment in which complete dehydration was observed when DCPD was held isothermally at 60◦ C for 5 days in a sealed glass capillary. In solid dosage forms, if the water released by the dehydration of DCPD is sorbed by the other formulation components, there will be no change in the total water content. However, the released water, in light of its mobility, may bring about undesirable physical and chemical changes. This is the subject of a current investigation in our laboratory.
1465
REFERENCES
We gratefully acknowledge the financial support from Dane O. Kildsig Center for Pharmaceutical Processing Research. Parts of this work were carried out in the Characterization Facility, University of Minnesota, which receives partial support from the National Science Foundation through the Materials Research Science and Engineering Centers program.
1. Debnath S Suryanarayanan R. 2004. Influence of processinginduced phase transformations on the dissolution of theophylline tablets. AAPS PharmSciTech 5:39–49. 2. Govindarajan R Suryanarayanan R. 2006. Processing- induced phase transformations and their implications on pharmaceutical products quality. In Polymorphism; Hilfiker R, Ed., Weinheim, Germany: Wiley-VCH, pp 333–364. 3. Sakata Y, Shiraishi S, Takayama K, Otsuka M. 2006. Effect of pulverization and dehydration on the pharmaceutical properties of calcium lactate pentahydrate tablets. Colloids Surf B Biointerfaces 51:149–156. 4. Otsuka M, Kaneniwa N, Otsuka K, Kawakami K, Umezawa O. 1993. Effect of tableting pressure and geometrical factor of tablet on dehydration kinetics of theophylline monohydrate tablets. Drug Dev Ind Pharm 19:541–557. 5. Moreton R. 2000. Calcium phosphate dibasic dihydrate. In Handbook of pharmaceutical excipients; Kibbe AH, Ed. London, UK: American Pharmaceutical Association, pp 63–67. 6. Moreton R. 2000. Calcium phosphate dibasic anhydrous. In Handbook of pharmaceutical excipients; Kibbe AH, Ed. London, UK: American Pharmaceutical Association, pp 60–62. 7. Schmidt PC, Herzog R. 1993. Calcium phosphates in pharmaceutical tableting. 1. Physico-pharmaceutical properties. Pharm World Sci 15:105–115. 8. Schmidt PC, Herzog R. 1993. Calcium phosphates in pharmaceutical tableting. 2. Comparison of tableting properties. Pharm World Sci 15:116–122. 9. Gohel MC, Parikh RK, Brahmbhatt BK, Shah AR. 2007. Preparation and assessment of novel coprocessed superdisintegrant consisting of crospovidone and sodium starch glycolate. AAPS PharmSciTech 8:E63–E69. 10. Landin M, Casalderrey M, Martinez-Pacheco R, Gomez-Amoza JL, Souto C, Concheiro A, Rowe RC. 1995. Chemical stability of acetylsalicylic acid in tablets prepared with different particle size fractions of a commercial brand of dicalcium phosphate dihydrate. Int J Pharm 123:143–144. 11. Duff EJ. 1971. Orthophosphates. XIII. Thermal decomposition of secondary calcium orthophosphate (CaHPO4 ) and secondary calcium orthophosphate dihydrate (CaHPO4 •2H2 O). J Appl Chem Biotechnol 21:233–235. 12. Dugleux P, De Sallier Dupin A. 1967. Kinetic studies of the dehydration of dicalcium orthophosphate. III. Evaluation of datascheme of reaction mechanism. Bull Soc Chim Fr 3:978–982. 13. Dugleux P, De Sallier Dupin A. 1967. Kinetic studies of the dehydration of dicalcium orthophosphate. II. Dehydration of calcium hydrogen phosphate dihydrate in an anhydrous or slightly wet atmosphere. Bull Soc Chim Fr 3:973–977. 14. Landin M, Rowe RC, York P. 1994. Structural changes during the dehydration of dicalcium phosphate dihydrate. Eur J Pharm Sci 2:245–252. 15. Landin M, Fontao MJ, Martinez-Pacheco R. 2005. A comparison of trehalose dihydrate and mannitol as stabilizing agents for dicalcium phosphate dihydrate based tablets. Drug Dev Ind Pharm 31:249–256. 16. Horhota ST, Burgio J, Lonski L, Rhodes CT. 1976. Effect of storage at specified temperature and humidity on properties of three directly compressible tablet formulations. J Pharm Sci 65:1746–1749. 17. Lausier JM, Chiang C-W, Zompa HA, Rhodes CT. 1977. Aging of tablets made with dibasic calcium phosphate dihydrate as matrix. J Pharm Sci 66:1636–1637. 18. Miyazaki T, Sivaprakasam K, Tantry J, Suryanarayanan R. 2009. Physical characterization of dibasic calcium phosphate dihydrate and anhydrate. J Pharm Sci 98:905–916. 19. Han J, Suryanarayanan R. 1997. Applications of pressure differential scanning calorimetry in the study of
DOI 10.1002/jps
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
CONCLUSIONS Dibasic calcium phosphate dihydrate, a widely used tablet diluent, dehydrates over pharmaceutically relevant timescales and conditions. The dehydration behavior is unusual in that it is accelerated in the presence of water vapor. The increase in dehydration rate with the increase in water vapor pressure can be rationalized if we consider the formation of a poorly crystalline anhydrate layer on the surface of dehydrating particles of DCPD. “Dry” conditions favor existence of the poorly crystalline layer and impede dehydration. On the contrary, in the presence of water vapor, recrystallization occurs, creating cracks and channels and facilitating removal of water. In a well-packaged solid dosage form, any water released upon initial dehydration is likely to be “trapped” and can accelerate further dehydration with potential implications on product attributes and performance.
ACKNOWLEDGMENTS
1466
20. 21.
22. 23.
KAUSHAL, VANGALA, AND SURYANARAYANAN
pharmaceutical hydrates. I. Carbamazepine dihydrate. Int J Pharm 157:209–218. Suryanarayanan R, Mitchell AG. 1986. Phase transitions of calcium gluceptate. Int J Pharm 32:213–221. Willart JF, De Gusseme A, Hemon S, Descamps M, Leveiller F, Rameau A. 2002. Vitrification and polymorphism of trehalose induced by dehydration of trehalose dihydrate. J Phys Chem B 106:3365–3370. Zhu H, Grant DJW. 2001. Dehydration behavior of nedocromil magnesium pentahydrate. Int J Pharm 215:251–262. Smith ML, Topley B. 1932. Heats of formation of the hydrates of manganese oxalate. J Chem Soc 1977–1979.
JOURNAL OF PHARMACEUTICAL SCIENCES, VOL. 100, NO. 4, APRIL 2011
24. Topley B, Smith ML. 1935. Kinetics of salt-hydrate dissociations: MnC2 O4 .2H2 O = MnC2 O4 + 2H2 O. J Chem Soc 57:321–325. 25. Seto Y, Sato H, Masuda Y. 2002. Effect of water vapor pressure on thermal dehydration of lithium sulfate monohydrate. Thermochim Acta 388:21–25. 26. Bertrand G, Lallemant M, Watelle-Marion G. 1974. Abnormal variation of the rate of decomposition of a solid. I. Dehydration of salt hydrates. J Inorg Nucl Chem 36:1303–1309. 27. Galwey A, Brown M. 1999. Thermal decomposition of ionic solids. Amsterdam, the Netherlands: Elsevier Science BV, pp 217–268.
DOI 10.1002/jps