Solubility and Speciation of Ketoprofen and Aspirin in Supercritical

Article pubs.acs.org/jced

Solubility and Speciation of Ketoprofen and Aspirin in Supercritical CO2 by Infrared Spectroscopy Mathilde Champeau,†,‡ Jean-Michel Thomassin,‡ Christine Jérôme,‡ and Thierry Tassaing*,† †

Institut des Sciences Moléculaires, UMR 5255 CNRS- University of Bordeaux, 351, Cours de la Libération, F-33405 Talence Cedex, France ‡ Department of Chemistry, Centre for Education and Research on Macromolecules (CERM), University of Liège, Sart-Tilman B6A, 4000 Liège, Belgium S Supporting Information *

ABSTRACT: The solubility of ketoprofen and aspirin in subcritical and supercritical CO2 was measured using FTIR absorption spectroscopy in the large range of temperature of 298.2−353.2 K and pressure of 5−35 MPa. The evolution of the solubility of both active pharmaceutical ingredients (APIs) was fitted using the Chrastil’s equation. In addition, the speciation of both APIs in monomeric and dimeric forms was explored by analyzing the characteristic carbonyl stretching vibrations of the carboxylic acid functions assigned to the dimers and monomers, respectively. Moreover, the evolution of the dimerization constant K of the two drugs as a function of the temperature and the pressure of scCO2 has been reported.

■

INTRODUCTION The pharmaceutical industry has gained a growing interest for supercritical CO2 (scCO2) due to the possibility of processing thermosensitive compounds and of recovering materials free of any solvent residues.1 ScCO2 is used for various processes such as the extraction of natural active pharmaceutical compounds (APIs) from roots, leaves, or seeds, the micronization of drugs, the impregnation or encapsulation of drug into a polymeric matrix,2−6 and so forth. For all these processes, the accurate knowledge of the solubility of APIs in scCO2 is required in order to design and optimize the processes. Solubility is defined as the mole fraction of drug solubilized in CO2 and can be adjusted by tuning pressure and temperature. It can be enhanced by a well-selected cosolvent that is generally ethanol or acetone because CO2 is a poor solvent for numerous hydrophilic and polar substances.7−10 Numerous studies have investigated the solubility of various APIs in subcritical and supercritical CO2. The solubility of these compounds has been published as review papers.11,12 Among them, Gupta and Shim have reported the solubility of 783 compounds that were published between the 1960s and 2004.13 More recently, Skerget et al. reviewed the solubility of solid compounds that were investigated between 2005 and 2010.14 The investigated APIs included anti-inflammatories, antibacterials, anticancers, βblockers, antipsychotics, antiglaucomas, anti-HIV, hormones, fibrates, vitamins, and so forth. To investigate the solubility of these compounds in CO2, various techniques have been used, such as gravimetric, HPLC, UV, FTIR, and cloud point determination.11,12,15 FTIR has the advantage of measuring the solubility, as well as the speciation of the molecules,16 by analyzing the characteristic bands of the solubilized molecules. © XXXX American Chemical Society

The dimerization of a molecule is likely to impact its diffusion in processes such as extraction or impregnation.17 This work is part of a project that aims at investigating the scCO2 impregnation process of semicrystalline polymers with ketoprofen and aspirin, two nonsteroid anti-inflammatory drugs bearing carboxylic acid functions (Figure 1).18,19 The solubility of ketoprofen and aspirin has been previously investigated but in a narrow range of temperature up to 338.2 and 328.2 K, respectively.7,20−25 The present work aims at determining their solubility in scCO2 in a larger range of temperature and pressure compared to what has been already published in the literature. The solubility of ketoprofen and aspirin in CO2 was measured at 298, 313.2, 328.2, and 353.2 K using FTIR absorption spectroscopy for pressure in the range 0.1−35 MPa. The evolution of the solubility was explained in terms of temperature and density of CO2. Then, the densitybased model Chrastil’s equation enabled to obtain the solubility at 343.2 and 363.2 K. Finally, because the carboxylic acids can be solubilized both in monomeric and dimeric forms their speciation was investigated as a function of temperature and pressure.

1. EXPERIMENTAL DETAILS 1.1. Materials. The information for each chemical samples are given in Table 1. Ketoprofen was used as received because the particles were thin enough to ensure a good solubilization of ketoprofen in scCO2, whereas acetylsalicylic acid was finely Received: September 22, 2015 Accepted: December 25, 2015

A

DOI: 10.1021/acs.jced.5b00812 J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data

Article

Figure 1. Chemical structure of (a) ketoprofen and (b) aspirin.

the cell. Two thermocouples were used, the first one located close to a cartridge heater for the temperature control and the second one close to the sample area to measure the temperature of the sample with a standard uncertainty u(T) = 1 K. The cell was connected via a stainless steel capillary tube to a manual pump purchased from TOP Industrie which allows the pressure to be raised up to 50 MPa with a standard uncertainty u(P) = 0.1 MPa and a relative standard uncertainty ur(P) = 0.003. The stabilization of the operating conditions was controlled by recording several consecutive spectra. The experiments were performed between 298.2 and 353.2 K and in a pressure range from 0.1 to 35 MPa. 1.3. Experimental Procedure. First, the drug powder was placed in the bottom of the cell in a cork that was then mounted on the cell (Figure 2). The powder was well below the incoming infrared beam such that the CO2 phase could be analyzed. Then, the cell was heated up to the required temperature followed by the slow addition of CO2 to reach the desired pressure in order to avoid any spread of the drug powder inside the cell. The system was kept under isobaric and isothermal conditions for a period between 10 and 30 min. During the stabilization of the operating conditions, consecutive spectra were recorded every 5 min. The equilibrium was considered to be achieved when no changes in the spectral bands were noticed. Once the equilibrium was reached, the pressure was raised to a higher value. The mixture was constantly homogenized during the experiment using a magnetically driven stirrer disposed into the cork. The stirring speed was optimized to avoid the powder to be spread into the cell. For solubility measurements, a large amount of drug was placed into the cell (∼30 mg), that is, in excess in order to allow saturation of CO2 with drug in all the studied conditions. 1.4. Infrared Absorption Spectra. The attenuated total reflectance-infrared (ATR-IR) spectra of aspirin and ketoprofen are presented in Figure 3. Both drugs display characteristic

Table 1. Information for the Chemical Samples chemical name carbon dioxide N45 aspirina ketoprofenb a

initial mole fraction purity

purification method

Air Liquide

99.95%

none

SigmaAldrich SigmaAldrich

≥ 99.0%

none

≥ 98.0%

none

source

2-(acetoxy)benzoic acid, that is, acetylsalicylic acid. benzoylphenyl)propanoic acid.

b

2-(3-

ground to form a powder with finer particles. The chemical structures of ketoprofen and aspirin are presented in Figure 1. 1.2. Experimental Infrared Setup. The infrared absorption measurements were performed using a ThermoFisher 6700 FTIR spectrometer equipped with a globar source, a KBr/ Ge beamsplitter and a deuterated triglycine sulfate (DTGS) detector that allows investigating the spectral range between 600 and 7000 cm−1. Single beam spectra recorded with a 2 cm−1 resolution were obtained after the Fourier transformation of 60 accumulated interferograms. A picture and a schematic representation of the experimental setup are presented in Figure 2. The infrared absorption experiments were performed using an in-house built stainless steel cell26 equipped with four cylindrical windows, two silicon windows with a variable path length between 0.412 and 2.395 cm for the infrared absorption measurements and two other sapphire windows for direct observation of the solution. The seal was obtained using the unsupported area principle. The windows were positioned on the surface of a stainless steel plug with a 100 μm Kapton foil placed between the window and the plug to compensate for any imperfections between the two surfaces. Teflon O-rings were used to ensure the seal between the plug and the cell body. The cell was heated using cartridge heaters disposed in the body of

Figure 2. Picture and schematic diagram of the high-pressure optical cell. B



Article

group of ketoprofen in its monomeric form and in its linear dimeric form. In order to support the assignment of the νCO stretching modes of the different ketoprofen species (monomer, linear dimer, and cyclic dimer), we performed density functional theory (DFT) calculations on ketoprofen monomer and the linear and cyclic form of ketoprofen dimer using the Gaussian 2009 package.27 DFT calculations of geometry, energies, and vibrational frequencies were carried out with the M062X functional using 6-31G basis sets. The optimized minimum energy structures are presented in Figure 5. Thus, we found that the calculated wavenumbers (without any scaling factor) for the νCO stretching vibration of the monomer, the linear dimer and the cyclic dimer were at about 1791, 1770, and 1743 cm−1, respectively. Even if the calculated wavenumbers do not have the same absolute value as the experimental ones, they are in the following order: νCO monomer > νCO linear dimer> νCO cyclic dimer; which supports the assignment for the experimental spectrum of ketoprofen solubilized into scCO2 (Figure 4b). Each ketoprofen molecule has a distinct peak in the νCO spectral region that can be used to determine the proportion of monomer, linear dimer, and cyclic dimer of ketoprofen in scCO2 as a function of pressure and temperature. The dissociation of dimers to form monomers was previously observed for various organic acids once solubilized in scCO2.28,29 However, few papers have reported the formation of linear dimeric form. Bell et al. observed this specie for trifluoroacetic acid in CO2 and, contrary to us, they found that the peak corresponding to the linear dimeric form appears at lower wavenumbers compared to the cyclic dimeric form.16 Because the νCO bands at 1763, 1742, and 1716 cm−1 are dependent on the speciation of ketoprofen, they could not be used to determine the solubility because the speciation is likely to evolve with temperature and pressure. Therefore, the bands centered at 1672 (νCO ketone), 1605 (νCC phenyl), and 1585 cm−1 (νCC phenyl) were selected to measure the solubility of ketoprofen, depending on their saturation or not. To determine the dimerization constant, the bands centered at 1763 (ν CO monomer ), 1742 (ν CO linear dimer ), and 1716 cm −1 (νCO cyclic dimer) were used. The peak heights were used instead of integrated area because the baseline choice produces larger errors when the integrated area method is used. Because the bands at 1763 and 1742 cm−1 overlapped, the height of these peaks could not be determined directly on the spectra. The deconvolution technique was applied to identify the contribution of these two peaks in the region 1725−1800 cm−1 by fitting the peak with a Gaussian−Lorentzian profile. 1.4.b. Aspirin Spectra. Similarly as previously done for ketoprofen, the characteristic peaks of crystallized aspirin were assigned (Figure 6). The assignments of the peaks have been reported elsewhere.30,31 The peak centered at 1780 cm−1 corresponds to the νCO stretching mode of the ester group, the peaks at 1750 and 1706 cm−1 are respectively attributed to the νCO stretching mode of the aspirin carboxyl group in its monomeric and cyclic dimeric forms, and the peaks at 1608 and 1584 cm−1 are both assigned to the νCC stretch of the phenyl group. Interestingly, the spectra of aspirin in scCO2 does not have a characteristic band of the νCO stretching mode of the aspirin carboxyl group in its linear dimeric form. The bands centered at 1780, 1608 and 1584 cm−1 were chosen to determine the solubility of aspirin into scCO2. Because these bands were saturated in some of the studied

Figure 3. Comparison of the ATR-IR spectra of aspirin and ketoprofen and an IR transmission spectrum of scCO2 in the highpressure optical cell (at T = 313.2 K P = 9 MPa).

bands in the range of 400 and 1800 cm−1 as well as a broad peak centered at 3000 cm−1. The IR spectrum of scCO2 has no contributions that overlap with the characteristic peaks of ketoprofen and aspirin in the range 1550−1800 cm−1 as shown in Figure 3. Therefore, the characteristic peaks of ketoprofen and aspirin in the range of 1550−1800 cm−1 have been selected for the present study. 1.4.a. Ketoprofen Spectra. The spectral range of 1550− 1800 cm−1 of the ATR-IR spectrum of ketoprofen powder (i.e., crystallized) is shown in Figure 4a. The peak centered at 1695

Figure 4. Comparison of the IR spectra of (a) ketoprofen powder and (b) ketoprofen solubilized in scCO2 (T = 313.2K P = 15 MPa).

cm−1 corresponds to the νCO stretching mode of the carboxylic acid group of ketoprofen molecules organized in a crystalline structure, two ketoprofen molecules forming cyclic dimers by hydrogen bonding. The peak at 1655 cm−1 is assigned to the νCO stretching mode of the ketone group and the peaks between 1560 and 1610 cm−1 are characteristic to the νCC stretching mode of the phenyl group. Once ketoprofen is solubilized in scCO2, some changes can be noticed (Figure 4b). The bands at 1716 and 1672 cm−1 correspond to the νCO stretching mode of the carboxyl group of ketoprofen in its cyclic dimeric form and to the νCO stretching mode of the ketone group, they are shifted to higher wavenumber compared to the bands at 1695 and 1655 cm−1 in crystallized ketoprofen due to the modification of the molecular environment. Moreover, two new bands appear at 1763 and 1742 cm−1 that are assigned to the νCO stretching modes of the carboxyl C



Article

Figure 5. Optimized structures of ketoprofen (a) monomer, (b) cyclic dimeric form, and (c) linear dimeric form. (d) Calculated IR spectrum in the CO stretching region of each optimized structure.

conditions, two other peaks centered at 1113 and 1132 cm−1 that did not saturate were selected (see Supporting Information Section a).

2. DATA PROCESSING 2.1. Solubility of Ketoprofen and Aspirin in scCO2. Figure 7 illustrates the evolution of the characteristic peaks of ketoprofen and aspirin with an increase in pressure. The drug solubility in CO2 could be calculated for each operational condition (P, T). The Beer−Lambert law was applied to the characteristic peak i of the drug to determine the concentration of drug Cdrug,i solubilized in CO2:

Figure 6. Comparison of the IR spectra of (a) aspirin powder and (b) aspirin solubilized in scCO2 (T = 313.2 K P = 13 MPa).

Cdrug, i =

Ai εil

(1)

Figure 7. Evolution of the IR spectra as a function of the pressure of (a) ketoprofen in scCO2 at T = 313.2 K and (b) of aspirin in scCO2 at T = 298.2 K. D



Article

Figure 8. (a) Solubility of ketoprofen in scCO2 at 313.2 K and (b) solubility of aspirin in scCO2 at 328.2 K as a function of the pressure. Data calculated with the three selected characteristic peaks and comparison with the literature data.

where Ai is the absorbance of the selected peak, εi is the molar extinction coefficient of the peak (L·mol−1·cm−1), and l is the path length (cm). The molar extinction coefficients εi of the characteristic peaks of ketoprofen and aspirin were determined (see Supporting Information Section b). In each condition, the concentration of drug Cdrug was considered to be the average of the concentration values Cdrug,i obtained with the different peaks i. The solubility y was then calculated in mole fraction for further comparison with literature data, the concentration of CO2 being known from the literature.32 y=

2.3. Semiempirical Correlation. In order to predict the solubility of the drugs at different temperature and pressure, a model based on a semiempirical density-based correlation was applied. The Chrastil model was selected among the existing semiempirical models because it is widely used, it is easy to implement, and generally it models well the solubility of solutes in subcritical and supercritical CO2.34−36 This model considers that one molecule of solute A associates with c molecules of CO2 (notes B) to create a complex ABc. This model relates the solubility of the solute to the density of CO2 and to the temperature as follows37 b (3) T −3 where S is the solubility of the solute in CO2 (kg·m ), ρ is the density of pure CO2 (kg·m−3), T is the absolute temperature (K), and a, b, and c are three adjustable parameters. By fitting the model to experimental solubility data, the three parameters a, b, and c for ketoprofen and aspirin can be determined. In order to estimate the correlation of the Chrastil model with the experimental data, the average absolute relative deviations (%AARD) between the calculated and the experimental solubility data can be estimated using the following equation ln S = a + c ln(ρ) +

Cdrug Cdrug + CCO2

(2)

Note that some peaks observed in Figure 7 are saturated at some pressures and could not be used for our calculations. 2.2. Validation of the Method by Comparison with the Literature Data. In order to validate our method, our results were compared to the data available in the literature. The solubility of ketoprofen at 313.2 K in the pressure range from 9.5 to 35 MPa was calculated from the three selected peaks (1585, 1605, and 1672 cm−1) and is reported in Figure 8a. The characteristic peaks at 9 MPa were too weak to be used for our purpose. In Figure 8a, the solubility is compared to the results obtained by Stassi et al.,21 by Weinstein et al.,23,33 and by Macnaughton et al. at 312.5 K.24 The results obtained with the three peaks appear to be slightly higher than the one of Macnaughton et al. However, our results are in good agreement with the results of Stassi et al. and with the two studies of Weinstein et al. hence validating the calculated values of the molar extinction coefficients. Good agreement was also observed for the solubility of aspirin calculated from the four selected peaks with the data reported by Huang et al.7 and Ravipaty et al.25 at 328.2 K (Figure 8b). Judging from Figure 8, the present method enabled to obtain data in good agreement with the literature data obtained either by a dynamic approach7,21,25 or by the cloud point technique.23 For the determination of solubility, the concentration of drug Cdrug was then considered to be the average of the concentration values Cdrug,i obtained with the three different peaks. We have evaluated a relative standard uncertainty ur(Cdrug) = 0.05 on the solubility data using our setup and our data processing.

%AARD =

100 n

n

∑

|yiexp − yicalc | yiexp

i=1

(4)

data points, yexp is i point i, and ycalc i

where n is the number of the experimental solubility of the drug for is the calculated solubility of the drug for point i. 2.4. Speciation of Drug. Ketoprofen and aspirin being carboxylic acids can not only be solubilized in scCO2 in their monomeric form but also in their cyclic dimeric form as presented in Figures 4 and 6. Ketoprofen can even be solubilized in its linear dimeric form. We aimed at determining the evolution of the ratio dimer/monomer with pressure and temperature by investigating the evolution of the dimerization constant K defined as follows 2 × monomer ↔ dimer

K=

Cdimer Cmonomer 2

(5)

with Cdimer = Ccyclic dimer + C linear dimer E

(6) DOI: 10.1021/acs.jced.5b00812 J. Chem. Eng. Data XXXX, XXX, XXX−XXX


Article

where Cmonomer is the concentration of monomeric form, Cdimer is the concentration of dimeric form (cyclic and linear) (in mol L−1), Ccyclic dimer is the concentration of cyclic dimeric form, and Clinear dimer is the concentration of linear dimer form. The concentration of drug solubilized in scCO2 corresponds to the sum of the concentration of the drug in its monomeric form Cmonomer and in its dimer form (2 × Cdimer) Cdrug = Cmonomer + 2 × Cdimer

eq 10 to the results obtained at 353.2 K. The molar extinction coefficient of ketoprofen in monomeric form was found to be ε1763,monomer = 432.9 L·mol−1·cm−1

The concentration Cdimer was calculated as follows using eq 7 Cdimer =

(11) 2 The dimerization constant K was finally calculated using eqs 5 after calculation of the concentrations Cmonomer and Cdimer.

(7)

When the linear dimeric form does not exist (Clinear dimer = 0), eq 7 is simplified as follows C = Cmonomer + 2 × Ccyclic dimer

3. RESULTS AND DISCUSSION 3.1. Experimental Solubility Results. The solubility measurements of ketoprofen and aspirin in subcritical and supercritical CO2 were performed at 298.2, 313.2, 328.2, and 353.2 K in the pressure range from 5 to 35 MPa. The solubility data of ketoprofen and aspirin are tabulated in Table 2 and 3,

(8)

The concentrations Cmonomer and Ccyclic dimer can be determined applying the Beer−Lambert law: Cmonomer =

A monomer εmonomerl

Ccyclic dimer =

Acyclic dimer εcyclic dimerl

(9)

Table 2. Solubility of Ketoprofen in scCO2 (y; mole fraction of ketoprofen)a

As previously proposed by Bell et al.,16 the molar extinction coefficient εmonomer was determined by combining eqs 8 and 9 applied in the case only when the cyclic dimer form exists. The following equation is obtained Cdrugl A monomer

=

1 εmonomer

Plotting the curve

⎛ ⎞⎛ Acyclic dimer ⎞ 2 ⎟⎟⎜⎜ ⎟⎟ + ⎜⎜ ⎝ εcyclic dimer ⎠⎝ A monomer ⎠

Cdrugl A monomer

=f

(

Acyclic dimer A monomer

), the

factor

Cdrug − Cmonomer

T (K)

P (MPa)

Ketoprofen solubility (105 y)

298.2

7.2 10 12.6 18 25 35 9.5 10 13 15 20 25.8 30 35 13 17.5 21.5 25 30 35 17 20 25 30 35

2.07 3.24 4.20 5.74 7.70 9.74 1.07 1.67 4.18 5.70 8.96 12.30 13.64 15.20 1.35 7.26 11.68 16.12 22.41 27.14 10.36 22.25 48.20 72.21 95.11

(10) 1 εmonomer

313.2

could be obtained since it corresponds to the intercept of the curve. The molar extinction coefficient of aspirin in monomeric form was thus determined from the value of the intercept of the fitting curve (Supporting Information Section c) ε1750,monomer = 862.1 L·mol−1·cm−1 328.2

Concerning ketoprofen, attention had to be paid to the presence of the linear dimeric specie. Interestingly, the band corresponding to the νCO stretching vibrations of the linear dimeric form showed a decrease in its absorbance with increasing the temperature and this band even disappeared at 353.2 K (Figure 9). Consequently, the molar extinction coefficient ε1763, monomer was determined by directly applying

353.2

a

Standard uncertainties u are u(T) = 1.0 K, u(p) = 0.1 MPa, ur(y) = 0.05.

respectively. The solubility curves obtained for ketoprofen and aspirin are also reported in Figures 10 and 11, respectively, as a function of the pressure (a) and as a function of the density of CO2 (b). In Figure 9, the results obtained by Sabegh et al. at T = 338.2 K are also plotted22 and this isotherm is found to be in between our isotherms at 328.2 and 353.2 K, which is concordant with the increase of the solubility with the temperature at fixed CO2 density. Moreover, the shapes of the isotherms of our data and those of Sabegh et al. are similar. Note that in the case of aspirin at 328.2 and 353.2 K, the data at 35 MPa are not presented because none of the characteristic peaks could be exploited at 30 and 35 MPa (all saturated).

Figure 9. Evolution of the νCO stretching vibration of the carboxyl group in linear dimeric form of ketoprofen solubilized in scCO2 at 15− 20 MPa at different temperatures of (a) 313.2, (b) 328.2, and (c) 353.2 K. F



Article

solvation power. Under isothermal conditions, the solubility of both drugs increases with pressure due to the increase in the density of CO2, as it has been widely reported in the literature.2,24,38,39 While increasing the CO2 density, the distance between the molecules of drug and CO2 decreases and the solvating power of CO2 is enhanced which explains the increase in the drug solubility. Under isobaric conditions, the effect of temperature is more complex and depends on the working pressure. As one can see in Figures 10a and 11a, increasing the temperature at a low pressure decreases the solubility, whereas it increases the solubility at a higher pressure. The effect of temperature changes until the so-called crossover pressure is attained.40 This crossover pressure is determined at the cross-section of the four isotherms and is in the range 12.5−17.5 MPa for aspirin and 15−17.5 MPa for ketoprofen. The value of the crossover pressure of ketoprofen is concordant with the value of 16.5 MPa reported by Macnaughton et al.24 This complex phenomenon can be explained by the competition of two factors. On one hand, the density of CO2 decreases with increasing the temperature that is unfavorable for the solubility. On the other hand, increasing the temperature enables to increase the sublimation pressure of the solute that is a favorable parameter. Below the crossover pressure, the effect of temperature is predominant, whereas it is the sublimation pressure above the crossover pressure. 3.2. Semiempirical Correlation. By fitting the Chrasil’s model to the experimental solubility data, the three parameters a, b, and c for ketoprofen and aspirin were determined (Table 4). Then, the solubility was calculated using Chrastil’s equation for the different conditions and at 343.2 and 363.2 K because these data were necessary for our investigation of the scCO2 impregnation process of semicrystalline polymers with these drugs.19 The results are compared to the experimental data in Figure 12. The Chrastil’s model applied to aspirin is able to correlate the experimental results in a better extent than for ketoprofen. Only the calculated results at 298.2 K do not fit the experimental data. The high %AARD values obtained by taking into account the data at all the temperatures can be explained by the limitation of the Chrastil’s equation. It has been previously noticed that Chrastil’s equation is not accurate over a wide range of temperature because of the wide range of solubility. Tabernero et al. reported the %AARD of various pharmaceutical compounds for which the solubilities were modeled by

Table 3. Solubility of Aspirin in scCO2 (y; Mole Fraction of Aspirin)a T (K)

P (MPa)

aspirin solubility (105 y)

298.2

7.5 10 12.5 18.5 25 30 35 9 9.5 10 13 15 20 25 30 35 11 12 13 15 17.5 20 24 28.5 12 13 15 17 20 24.8 30

7.56 9.73 11.22 14.03 16.87 18.10 19.88 1.95 3.41 4.79 9.18 11.31 17.04 18.97 19.99 22.40 2.21 4.99 7.62 13.07 18.76 23.54 29.07 37.33 4.05 5.74 11.21 18.60 37.83 64.03 91.95

313.2

328.2

353.2

a

Standard uncertainties u are u(T) = 1.0 K, u(p) = 0.1 MPa, ur(y) = 0.05.

The solubility of ketoprofen varies between 1.1 × 10−5 and 9.5 × 10−4 at T = 353.2 K, P = 35 MPa and the solubility of aspirin varies between 1.9 × 10−5 and 9.2 × 10−4 at T = 353.2 K, P = 30 MPa. Independent of the temperature, both ketoprofen and aspirin are insoluble at a pressure below 7.5 MPa since CO2 is not in its supercritical state in these conditions and thus has a low

Figure 10. Solubility of ketoprofen in CO2 at 298.2 K (green square); 313.2 K (pink triangle); 328.2 K (blue circle); 353.2 K (yellow right-facing triangle); and 338.2 K (violet square, data from Sabegh et al.) (a) as a function of the pressure and (b) as a function of the density of CO2. G



Article

Figure 11. Solubility of aspirin in CO2 at 298.2 K (green square); 313.2 K (pink triangle); 328.2 K (blue circle); 353.2 K (yellow right-facing triangle) (a) as a function of the pressure and (b) as a function of the density of CO2.

molecule of drug must evolve with the operational conditions as it will be discussed in Section 3.3. Therefore, the Chrastil’s model did not allow a good fitting for all the solubility data of ketoprofen and aspirin in a wide range of pressure and temperature. A better fitting should have been obtained using a model that takes into account a variation of the enthalpy of vaporization and of the association parameter with the operational conditions such as the Sparks model.42 3.3. Speciation of Drug. We aimed at investigating the evolution of the dimerization constant K with pressure and temperature. The dimerization constant K was calculated using eq 6 for each condition. The results obtained for ketoprofen and aspirin are compared in Figures 13 and 14, respectively, and plotted as a function of the pressure (a) and as a function of the density of CO2 (b). Whereas the density of CO2 is the parameter that enables us to explain the evolution, we also present the results as a function of the pressure because it is the parameter that is technically controlled. Comparing the results obtained with the two drugs, we can see that the dimerization constant of ketoprofen and aspirin are in the same range of order. At a fixed temperature, the dimerization constant decreases with increasing the pressure (i.e., increasing the density) for both drugs, which is due to the increase in the concentration of the monomer specie. Increasing the density of CO2, the CO2− carboxylic acid interactions are favored, which leads to a better stabilization of the monomer specie.16,28 A molecule of CO2

Table 4. Values of the Three Constants Used in Chrastil Model drug

a

b

c

%AARD

ketoprofen aspirin

−18.5 −16.4

−8796 −5 741

6.69 5.08

28.4 16.9

Chrastil’s equation, the highest values (up to 52.9%) were obtained for compounds that had a large range of solubility in the investigated temperature and pressure ranges.41 In fact, when temperature approaches the melting point of the solute, the solubility increases to a great extent and the enthalpy of vaporization and the change in the association parameter c must be taken into account to model the solubility, which is not the case of Chrastil’s equation.41,42 The melting temperatures of ketoprofen and aspirin under atmospheric pressure are 367.2 and 409.2 K, respectively, but are significantly decreased in scCO2. They were estimated to fall between 343 and 353 K for ketoprofen and between 373 and 383 K for aspirin according to the sticky appearance of the drug remaining in the reactor used for the experiments above these temperatures after depressurization.19 It has been confirmed by 1H NMR that the drugs are not decomposed in those conditions (see Supporting Information Section d). Therefore, the enthalpy of vaporization should not be considered to be constant. Moreover, the association parameter c, that is, the number of CO2 molecules that associates with one

Figure 12. Comparison of experimental (points) and calculated (lines) solubility thanks to the Chrastil’s model at various temperatures, as a function of the pressure for (a) ketoprofen and (b) aspirin. H



Article

Figure 13. Evolution of the dimerization constant of ketoprofen solubilized in scCO2 with (a) pressure and (b) density of CO2.

Figure 14. Evolution of the dimerization constant of aspirin solubilized in scCO2 with (a) pressure and (b) density of CO2.

4. CONCLUSIONS The solubility of ketoprofen and aspirin in subcritical and supercritical CO2 was measured using FTIR spectroscopy. Some characteristic peaks of each drug were identified. While aspirin is only solubilized in monomeric and cyclic dimeric forms, ketoprofen is also solubilized in its linear dimeric form. This linear dimeric form disappears at 353.2 K. The molar extinction coefficients of these peaks were calculated and their values were validated by comparing the solubility of each drug calculated with these coefficients with literature data. The solubility of the drugs was then determined at 298.2, 313.2, 328.2, and 353.2 K and at pressure range from 5 to 35 MPa. The results show that both drugs have a good solubility in CO2 ranging from 1.1 × 10−5 to 9.5 × 10−4 (mole fraction). The solubility of ketoprofen and aspirin can be reasonably considered as similar in the same experimental conditions and highly dependent on the pressure and temperature. Under isothermal conditions, the solubility increases with an increase in density of CO2. For pressures above the crossover pressure (15 MPa for aspirin and 17 MPa for ketoprofen), increasing the temperature leads to an increase in the solubility. The opposite effect is observed below the crossover pressure. The experimental data were then fitted using the Chrastil’s model, a semiempirical density-based model. The parameters of this model were fitted but poor correlation was obtained for both ketoprofen and aspirin (%AARD = 28.4 and 16.9%, respectively) because the Chrastil’s equation considers the enthalpy of vaporization and the change in the association parameter c to be constant even in a large range of pressure and temperature. Finally, the evolution of the dimerization constant

can interact with a carboxylic acid through Lewis acid/base as it has been proved by ab initio calculations43,44 and by spectroscopy.45 At a fixed density, the dimerization constant decreases when the temperature is increased as it has been already observed for other carboxylic acids solubilized in CO2.16 This can be explained by the fact that the hydrogen bonds between carboxylic acids are greatly weakened with increasing temperature. The dimerization of a molecule is likely to impact its diffusion in processes, such as supercritical CO2 assisted extraction and impregnation. Ngo et al. noticed that the scCO2 assisted impregnation of a polymer with a molecule that interacts strongly with the matrix can slow down the diffusion of this molecule into the polymer.17 In that case, they proposed to select molecules that dimerize in order to reduce the interactions and favor the diffusivity. According to the present results, selecting appropriate experimental conditions (i.e., low pressure and low temperature) can also favor the molecule− molecule interactions for a given molecule. Finally, the evolution of the dimerization constant K indicates that the drug−drug and CO2−drug interactions change with pressure and temperature. Therefore, the association parameter c used in the Chrastil equation must evolve with the experimental conditions whereas it is considered to be constant in the Chrastil model. This can explain the failure of Chrastil model in correlating the solubility of ketoprofen and aspirin for the wide range of pressure and temperature applied in this study. I



Article

(7) Huang, Z.; Lu, W. D.; Kawi, S.; Chiew, Y. C. Solubility of Aspirin in Supercritical Carbon Dioxide with and without Acetone. J. Chem. Eng. Data 2004, 49, 1323−1327. (8) Duarte, A. R. C.; Santiago, S.; De Sousa, H. C.; Duarte, C. M. M. Solubility of acetazolamide in supercritical carbon dioxide in the presence of ethanol as a cosolvent. J. Chem. Eng. Data 2005, 50, 216− 220. (9) Duarte, A. R. C.; Coimbra, P.; De Sousa, H. C.; Duarte, C. M. M. Solubility of flurbiprofen in supercritical carbon dioxide. J. Chem. Eng. Data 2004, 49, 449−452. (10) Besnard, M.; Tassaing, T.; Danten, Y.; Andanson, J.-M.; Soetens, J.-C.; Cansell, F.; Loppinet-Serani, A.; Reveron, H.; Aymonier, C. Bringing together fundamental and applied science: The supercritical fluids route. J. Mol. Liq. 2006, 125, 88−99. (11) Fonseca, J. M. S.; Dohrn, R.; Peper, S. High-pressure fluid-phase equilibria: Experimental methods and systems investigated (2005− 2008). Fluid Phase Equilib. 2011, 300, 1−69. (12) Dohrn, R.; Peper, S.; Fonseca, J. M. S. High-pressure fluid-phase equilibria: Experimental methods and systems investigated (2000− 2004). Fluid Phase Equilib. 2010, 288, 1−54. (13) Gupta, R. B.; Shim, J. J. Solubility in Supercritical Carbon Dioxide. CRC Press, Boca Raton, 2007. (14) Škerget, M.; Knez, Z.; Knez-Hrnčič, M. Solubility of solids in sub- and supercritical fluids: A review. J. Chem. Eng. Data 2011, 56, 694−719. (15) Oparin, R. D.; Idrissi, A.; Fedorov, M. V.; Kiselev, M. G. Dynamic and Static Characteristics of Drug Dissolution in Supercritical CO2 by Infrared Spectroscopy: Measurements of Acetaminophen Solubility in a Wide Range of State Parameters. J. Chem. Eng. Data 2014, 59, 3517−3523. (16) Bell, P. W.; Thote, A. J.; Park, Y.; Gupta, R. B.; Roberts, C. B. Strong Lewis Acid-Lewis Base Interactions between Supercritical Carbon Dioxide and Carboxylic Acids: Effects on Self-association. Ind. Eng. Chem. Res. 2003, 42, 6280−6289. (17) Ngo, T. T.; Liotta, C. L.; Eckert, C. A.; Kazarian, S. G. Supercritical fluid impregnation of different azo-dyes into polymer: In situ UV/Vis spectroscopic study. J. Supercrit. Fluids 2003, 27, 215− 221. (18) Champeau, M.; Thomassin, J. M.; Jerome, C.; Tassaing, T. In situ investigation of supercritical CO2 assisted impregnation of drugs into a polymer by high pressure FTIR micro-spectroscopy. Analyst 2015, 140, 869−879. (19) Champeau, M.; Thomassin, J.-M.; Tassaing, T.; Jerome, C. Drug Loading of Sutures by Supercritical CO2 Impregnation: Effect of Polymer/Drug Interactions and Thermal Transitions. Macromol. Mater. Eng. 2015, 300, 596−610. (20) Tsivintzelis, I.; Economou, I. G.; Kontogeorgis, G. M. Modeling the phase behavior in mixtures of pharmaceuticals with liquid or supercritical solvents. J. Phys. Chem. B 2009, 113, 6446−6458. (21) Stassi, A.; Bettini, R.; Gazzaniga, A.; Giordano, F.; Schiraldi, A. Assessment of solubility of ketoprofen and vanillic acid in supercritical CO2 under dynamic conditions. J. Chem. Eng. Data 2000, 45, 161− 165. (22) Sabegh, M. A.; Rajaei, H.; Esmaeilzadeh, F.; Lashkarbolooki, M. Solubility of ketoprofen in supercritical carbon dioxide. J. Supercrit. Fluids 2012, 72, 191−197. (23) Weinstein, R. D.; Muske, K. R.; Martin, S. A.; Schaeber, D. D. Liquid and supercritical carbon dioxide-assisted implantation of ketoprofen into biodegradable sutures. Ind. Eng. Chem. Res. 2010, 49, 7281−7286. (24) Macnaughton, S. J.; Kikic, I.; Foster, N. R.; Alessi, P.; Cortesi, A.; Colombo, I. Solubility of anti-inflammatory drugs in supercritical carbon dioxide. J. Chem. Eng. Data 1996, 41, 1083−1086. (25) Ravipaty, S.; Koebke, K. J.; Chesney, D. J. Polar mixed-solid solute systems in supercritical carbon dioxide: Entrainer effect and its influence on solubility and selectivity. J. Chem. Eng. Data 2008, 53, 415−421. (26) Chambon, P.; Cloutet, E.; Cramail, H.; Tassaing, T.; Besnard, M. Synthesis of core-shell polyurethane-polydimethylsiloxane particles

K of the two drugs in scCO2 as a function of temperature and pressure was investigated. The equilibrium between carboxylic acid dimer and monomer was explored by analyzing the characteristic bands νCO stretching vibration of the carbonyl function assigned to the dimer and monomer, respectively. At a given temperature, the constant K decreases by increasing the density of CO2. It is accounted by more favorable solute− solvent attractive interactions upon increasing the pressure that stabilizes the monomer specie. At a fixed density, the dimerization constant decreases when the temperature is increased as the hydrogen bond interactions between carboxylic acids are weakened with increasing the temperature. The dimerization constant K of ketoprofen and aspirin are similar at a given temperature and pressure.

■

ASSOCIATED CONTENT

S Supporting Information *

IThe Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jced.5b00812. IR spectrum of aspirin in scCO2, determination of the molar extinction coefficients εi of the characteristic peaks of ketoprofen and aspirin, and 1H NMR spectra of ketoprofen and aspirin after treatment in scCO2 at 413 K and 30 MPa for 5 h. (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Fax: +33540006994. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge the ‘‘Conseil Régional Aquitaine’’ for the Ph.D. fellowship of M. Champeau and support for working costs. This work was performed in the frame of the international doctoral school program “IDS-FunMat” supported by the European Community (ERASMUS MUNDUS Doctoral program). CERM is much indebted to IAP VII-05 “Functional Supramolecular Systems” (FS2). J.-M. T. is Logistic Collaborator by the FRS-FNRS.

■

REFERENCES

(1) Davies, O. R.; Lewis, A. L.; Whitaker, M. J.; Tai, H.; Shakesheff, K. M.; Howdle, S. M. Applications of supercritical CO2 in the fabrication of polymer systems for drug delivery and tissue engineering. Adv. Drug Delivery Rev. 2008, 60, 373−387. (2) Marceneiro, S.; Coimbra, P.; Braga, M. E. M.; Dias, A. M. A.; de Sousa, H. C. Measurement and correlation of the solubility of juglone in supercritical carbon dioxide. Fluid Phase Equilib. 2011, 311, 1−8. (3) Türk, M. Manufacture of submicron drug particles with enhanced dissolution behaviour by rapid expansion processes. J. Supercrit. Fluids 2009, 47, 537−545. (4) Foster, N.; Mammucari, R.; Dehghani, F.; Barrett, A.; Bezanehtak, K.; Coen, E.; Combes, G.; Meure, L.; Ng, A.; Regtop, H. L.; Tandya, A. Processing Pharmaceutical Compounds Using Dense Gas Technology. Ind. Eng. Chem. Res. 2003, 42, 6476−6493. (5) Kerč, J.; Srčič, S.; Knez, Z.; Senčar-Božič, P. Micronization of drugs using supercritical carbon dioxide. Int. J. Pharm. 1999, 182, 33− 39. (6) Champeau, M.; Thomassin, J. M.; Tassaing, T.; Jérôme, C. Drug loading of polymer implants by supercritical CO2 assisted impregnation: A review. J. Controlled Release 2015, 209, 248−259. J



Article

(44) Nelson, M. R.; Borkman, R. F. Ab initio calculations on CO2 binding to carbonyl groups. J. Phys. Chem. A 1998, 102, 7860−7863. (45) Kazarian, S. G.; Vincent, M. F.; Bright, F. V.; Liotta, C. L.; Eckert, C. A. Specific intermolecular interaction of carbon dioxide with polymers. J. Am. Chem. Soc. 1996, 118, 1729−1736.

in cyclohexane and in supercritical carbon dioxide used as dispersant media: a comparative investigation. Polymer 2005, 46, 1057−1066. (27) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, Revision A.2, Gaussian Inc., Wallingford CT, 2009. (28) Park, Y.; Gupta, R. B.; Curtis, C. W.; Roberts, C. B. Solvent effects on the self-association of formic acid in carbon dioxide and ethane. J. Phys. Chem. B 2002, 106, 9696−9700. (29) Ke, J.; Jin, S.; Han, B.; Yan, H.; Shen, D. Hydrogen bonding of some organic acid in supercritical CO2 with polar cosolvents. J. Supercrit. Fluids 1997, 11, 53−60. (30) Champeau, M.; Thomassin, J. M.; Jérôme, C.; Tassaing, T. In situ investigation of supercritical CO2 assisted impregnation of drugs into a polymer by high pressure FTIR micro-spectroscopy. Analyst 2015, 140, 869−879. (31) Boczar, M.; Wójcik, M. J.; Szczeponek, K.; Jamróz, D.; Ziba, A.; Kawałek, B. Theoretical modeling of infrared spectra of aspirin and its deuterated derivative. Chem. Phys. 2003, 286, 63−79. (32) NIST http://webbook.nist.gov/chemistry/ (accessed 09/2014). (33) Weinstein, R. D.; Muske, K. R.; Moriarty, J.; Schmidt, E. K. The Solubility of Benzocaine, Lidocaine, and Procaine in Liquid and Supercritical Carbon Dioxide. J. Chem. Eng. Data 2004, 49, 547−552. (34) Yamini, Y.; Tayyebi, M.; Moradi, M.; Vatanara, A. Solubility of megestrol acetate and levonorgestrel in supercritical carbon dioxide. Thermochim. Acta 2013, 569, 48−54. (35) Lee, C. A.; Tang, M.; Ho, S. L.; Chen, Y. P. Solubilities of chlormezanone, metaxalone and methocarbamol in supercritical carbon dioxide. J. Supercrit. Fluids 2014, 85, 11−16. (36) Rajaei, H.; Hezave, A. Z.; Lashkarbolooki, M.; Esmaeilzadeh, F. Representing experimental solubility of phenylephrine hydrochloride in supercritical carbon dioxide and modeling solute solubility using semi-empirical correlations. J. Supercrit. Fluids 2013, 75, 181−186. (37) Chrastil, J. Solubility of solids and liquids in supercritical gases. J. Phys. Chem. 1982, 86, 3016−3021. (38) Nalesnik, C. A.; Hansen, B. N.; Hsu, J. T. Solubility of pure taxol in supercritical carbon dioxide. Fluid Phase Equilib. 1998, 146, 315− 323. (39) Suleiman, D.; Antonio Estévez, L.; Pulido, J. C.; Garcia, J. E.; Mojica, C. Solubility of anti-inflammatory, anti-cancer, and anti-HIV drugs in supercritical carbon dioxide. J. Chem. Eng. Data 2005, 50, 1234−1241. (40) Chimowitz, E. H.; Pennisi, K. J. Process synthesis concepts for supercritical gas extraction in the crossover region. AIChE J. 1986, 32, 1665−1676. (41) Tabernero, A.; del Valle, E. M. M.; Galán, M. A. A comparison between semiempirical equations to predict the solubility of pharmaceutical compounds in supercritical carbon dioxide. J. Supercrit. Fluids 2010, 52, 161−174. (42) Sparks, D. L.; Hernandez, R.; Estévez, L. A. Evaluation of density-based models for the solubility of solids in supercritical carbon dioxide and formulation of a new model. Chem. Eng. Sci. 2008, 63, 4292−4301. (43) Raveendran, P.; Wallen, S. L. Cooperative C-H—— O hydrogen bonding in CO2-Lewis base complexes: Implications for solvation in supercritical CO2. J. Am. Chem. Soc. 2002, 124, 12590− 12599. K


Solubility and Speciation of Ketoprofen and Aspirin in Supercritical

Recommend Documents