Swelling Studies of Gelatin I: Gelatin Without Additives

Swelling Studies of Gelatin I: Gelatin Without Additives

Swelling Studies of Gelatin I: Gelatin Without Additives C.M. OFNER,111, AND HANS SCHOI-? Received August 30,1985, from the School of Pharmacy, Tem...

760KB Sizes 2 Downloads 92 Views

Swelling Studies of Gelatin I: Gelatin Without Additives C.M. OFNER,111,

AND

HANS SCHOI-?

Received August 30,1985, from the School of Pharmacy, Temple University, Philadelphia, PA Abstract 0 The swelling rate and the equilibrium swelling of gelatin (type B) were studied by casting warm gelatin solutions into films, cutting

them into short rectangular strips after gelation, drying them, and measuring the weight gain on immersion in buffer solutions as a function of time. The process variables investigated included concentration of the gelatin casting solutions, the thickness, drying conditions, age and residual moisture content of the film strips, the chemical nature and concentration of the buffers in the swelling solutions, and the temperature of these solutions at a constant pH of 7.0 (1.9 pH unit$ above the isoionic point). The swelling kinetics followed a second-order equation. The initial swelling rate and the equilibrium swelling of the amorphous portion of the gelatin strips (which was somewhat smaller than the total observed swelling) were calculated from a linearized form of the rate equation. 01 the factors investigated, the equilibrium swelling was increased most strongly when the temperature of the swelling solution was raised from 20 to 25 "C. Strip thickness was the predominant factor governing the rate of swelling, which was inversely proportional to the thickness. Conditions leading to slower drying and longer storage limes promoted more extensive crystallization, thereby increasing the density of the gelatin strips and reducing their swelling rate.

Even though gelatin is the most widely used polymer in pharmaceutical products, its physical properties have not been thoroughly investigated. A recent review' states t h a t "Collagen and its (semi-crystalline or) amorphous form, gelatin, have yet to be characterized definitely as pure polymeric substances in the bulk state." One important physical property, which strongly affects the behavior of gelatin under use conditions, is swelling in aqueous media. The following factors have been investigated individually: origin and method of manufacturing gelatin;2s3temperature and humidity during the preparation of film samples used for swelling meas~rements4.~ and their thickness;s and ionic strength' and temperature of the swelling solutions.6.6However, the results of these investigations cannot be compared with one another because they are not based on similar gelatin samples nor on comparable swelling techniques and measurements. The purpose of the present work was to investigate the influence of a broad range of factors on the kinetics of swelling of gelatin and on the maximum water uptake, employing a single batch of gelatin and a single method for measuring swelling. Swelling can be described in terms of rate and of maximum uptake a t equilibrium. Robinsone found empirically that a second-order rate equation described the rate of swelling in terms of the one-dimensional increase in thickness of gelatin films coated onto a solid support.6.@JOThe linear expression of Robinsons was adapted to the present method, based on the increase in weight of short rectangular strips of unsupported gelatin film, by replacing film thickness with the weight, W, of buffer solution absorbed per gram of anhydrous gelatin as a function of time t: t

-=A+BBt

W

790 / Journal of Pharmaceutical Sciences Vol. 75, No. 8, August 1986

19140.

Accepted for publication June 5, 1986.

where A and B are constants. Rearranging and differentiating results in:

-dW _ dt

A

(A

+ Bt12

For t --* 0, the initial rate of swelling is lIA, the reciprocal of the intercept in plots of NW versus t according ta eq. 1. The reciprocal of the slope, 1/B, equals W,, the maximum or equilibrium uptake, because at long times, Bt >> A. The units of A and B are (h * g gelatidg buffer solution) and (g gelatidg buffer solution), respectively. Among the other 35 equations tested was one with a square root and one with a logarithmic time dependence, as well as a first-order equation. None gave an approximately linear relation over the entire swelling period.

Experimental Section Materials-The gelatin was commercial type B, manufactured by alkaline treatment of decalcified bones by Kind & Knox, Sioux City, Iowa. It was USP grade, had an approximate bloom strength of 250 g, and contained 1.16 mmol of acidic amino acid residues and 0.96 mmol of basic residues per gram of dry gelatin." The swelling solution was buffered to pH 7.0 with 0.15 M buffer. The original buffer, triethanolamine hydrochloride, was replaced by Tris-HC1 because the latter has a greater buffer capacity. Ammonium acetate was finally ueed as the standard buffer in place of Tria-HC1to avoid a possible reaction between the cationic Tris and the carboxylate anions of the gelatin. The pH of the 0.15 M buffer solutions was adjusted to 7.0 by adding 1.0 M NaOH to triethanolamine hydrochloride and 1.0 M HCI to Tris. Ammonium acetate solutions required no adjustment. Water was doubly distilled and free of carbon dioxide. All chemicals were reagent grade except for hexadecane, which was practical grade. Moisture Determination of Gelatin-Even after storage under reduced pressure over a desiccant (anhydrous CaS04, "Drierite") for as long as three years, gelatin samples retained as much as 1.7% (w/w) moisture. The moisture content was determined by drying -500 mg of granules or pieces of the strips, weighed to the nearest mg, a t 105 "C for 18 '. 1h. A check for constant weight was made by heating for a n additional 4 h. Density Determination-Dried gelatin in the form of granules or pieces from cast strips was measured a t 25 2 0.2"C in a 25-mL pycnometer with hexadecane. Casting and Swelling Variable+"he variables investigated in the development of the standard casting and swelling procedure included: gelatin concentration of the casting solution (15, 20, and 25% w/w); dry strip thickness (0.17, 0.23, 0.28, and 0.29 mm); conditioning time of the cut strips while still in the gel state at 25 "C and 85% relative humidity (12 and 72 h); age (0.5,2,4-6, 11, and 39 months) and residual moisture (2.5 and 4.0%) of the stored strips prior to swelling; buffer employed in the swelling experiments (triethanolamine hydrochloride, Tris-HC1, and ammonium acetate); molarity of the buffer (0.10, 0.30, and 0.90 M);and temperature of the swelling solutions (20 and 25 "C). Preparation of Gelatin Strips-The four successive steps were: preparation of warm gelatin casting solutions; casting the solutions into films; cutting the films into short rectangular strips and conditioning them; and drying the strips to constant weight. Two hundred-gram batches of gelatin solution were prepared in a 1000-mL resin kettle whose cover contained standard-tapered openOO~-3549/86/08oO-0790$0 l.OO/O 6 7986, American PharmaceuticalAssociation

ings for a thermometer, a reflw condenser to minimize water loss, and a stirrer. A precision-ground glass rod was held in place by a matching glass sleeve which fitted snugly into the central opening of the cover. A stainless-steel stirrer shaft bearing a three-blade propeller was attached to the lower end of the glass rod. A drop of mineral oil provided lubrication between the glass rod and the sleeve. A reduced pressure of 40 mmHg, maintained by means of an aspirator connected to the top of the condenser, prevented the entrapment of air bubbles in the viscous gelatin solutions. The bottom half of the kettle was surrounded by a heating mantle. The standard procedure for preparing and casting gelatin solutions is described in detail because small deviations frequently resulted in large variations in strip properties. Ninety milliliters of ice water was placed in the resin kettle, followed by enough granular gelatin to give a concentration of 33 to 56% (w/w). ARer brief mixing, the stirring speed was reduced to the minimum rate of 25 rpm. Five minutes later, the heater was turned on to 55% of full scale of the variable transformer. The kettle was evacuated as the swollen gelatin granules, which had absorbed the water, melted to a thick, viscous solution. After 30 min from the start, when the temperature had risen to 58 "C, the heater and aspirator were turned off and air was admitted. Any gelatin adhering to the kettle walls was scraped off and returned to the solution. After 35 min from the start, a predetermined volume of warm 1.00 M NaOH required to adjust the pH to 7.0 was added. "he solution was mixed rapidly for a brief period, after which the stirring speed was again reduced to the minimum. The kettle walls were scraped periodically. After 55 min from the start, the balance of warmed water was added to complete the weight of the solution to 200 g. After 65 min, any water loss resulting from the repeated openings was determined by weight and replaced. After 85 min, the heating mantle was removed and the temperature dropped to 45 "C. After 90 min, when the solution had cooled to 40 "C, it was cast with an adjustable casting knife (No. 577, Gardner Laboratory, Inc., Bethesda, MD) onto a level glass plate equilibrated to 25 "C. The cast solution was stored for 2 h in a n incubator at 85 2 2% relative humidity and 25 "C. Short strips were cut to 2 x 5 cm size, placed on teflon-coated trays, and conditioned for 72 h in the incubator. If the strips were left to dry on the glass plate, the adhesion was so strong that shrinkage and detachment of the strips during drying often delaminated the glass, leaving flakes imbedded in the gelatin. The strips were then partially dried at ambient humidity for 3 h and, finally, dried to constant weight under reduced pressure over anhydrous CaSO,. The few strips which curled during conditioning, and those with air bubbles, were discarded. The average thickness of each strip was determined from five measurements to the nearest 0.01 mm with a Starrett thickness gauge (No. 25-481, L. S. Starrett Co., Athol, MA). For a swelling experiment employing five strips, the average strip thickness and its SD were therefore calculated from the 25 individual measurements. The dried strips measured 1.7 x 4.6 cm. Their thickness ranged from 0.16 to 0.54 mm, depending on the setting of the casting knife. The five strips selected for a single swelling experiment usually had the same thickness within 55%. The strips were stored at room temperature over anhydrous CaSO, (Drierite) for periods up to three years prior to the swelling experiments. The residual moisture content of each batch was determined before swelling experiments by drying a few strips a t 105 "C. These oven-dried strips were then discarded. Swelling Measurements-Each short rectangular strip was swollen in a separate 240-mL (8-oz) square amber glass bottle containing 200 mL of buffer solution. The capped bottles were placed in a constant temperature bath for the duration of the swelling experiments. Each swelling experiment employed three to seven strips. The strips were weighed every 15 min for the first hour and every 30 min for the following 3 h. Longer swelling studies included one measurement at up to 12 h, followed by a subsequent measurement every 24 h. The longest swelling times employed, 96 h, produced near-equilibrium uptake. Longer times weakened the strips excessively and occasionally led to microbial growth. At the appropriate time intervals, each strip was removed from the buffer solution with teflon forceps, briefly patted with lint-free cleaning tissues to remove the solution wetting its surface, weighed in a closed weighing bottle, and returned to its swelling solution. The total time a strip spent outside the swelling solution per measurement was <1 min.

At the completion of the swelling experiments, the pH of the buffer solutions changed <0.1 unit from the initial pH of 7.0. Swelling is expressed as grams of buffer solution absorbed per gram of dry gelatin. These numbers are identical with the grams of water absorbed per gram of dry gelatin within 1% for the 0.15 M ammonium acetate buffer. Determination of Gelatin Leaching-The strips lost small amounts of gelatin during the swelling experiments: two to three percent was leached out in 4 h, and a total of 57% in 96 h. The amount of gelatin dissolved as a function of time was determined with a modified micro Lowry protein assay.12 using three sets of strips. The first set (run 13) was conditioned for the standard 72 h at 25 "C and 85% relative humidity, and stored for 2 months prior to swelling. The second set (run 15) was conditioned for only 12 h; the third set (run 10) was stored for 39 months prior to swelling. When the buffers triethanolamine hydrochloride and Tris-HC1, which interfered with the gelatin assay, were used, the total amount of leached gelatin was determined at the end of the swelling experiments by drying the strips at 105°C for 18 h to constant weight. The weight of triethanolamine or Tris-HC1 dissolved in the buffer solution imbibed by the strips was subtracted from the strip weight. Compensation was made for the small amount of imbibed triethanolamine hydrochloride volatilized at the drying conditions.

Results and Discussion Gelatin Density-The factors affecting the density of gelatin were conditions during preparation of the strips, storage time at room temperature, and oven drying. The gelatin strips had higher densities (in units of g/cm3) than the original gelatin granules at comparable moisture contents. For instance, for oven-dried samples, the density of granular gelatin was 1.30, while strip densities ranged from 1.33 to 1.36. The density of strips stored over anhydrous CaS04 (Drierite) at room temperature aRer casting and conditioning increased with storage time. At a 2% moisture content, the highest density observed was 1.49 after 11 months, compared with a typical density of 1.36 at half that time or less. Oven drying reduced density as well as moisture content: a strip dried to 1.6%moisture over anhydrous CaS04 (Drierite) at room temperature had a density of 1.46; oven drying reduced it to 1.33. Leaching of Gelatin During Swelling-The three profiles of percent gelatin leached versus time (Table I) were nearly the same. Conditioning and storage times affected the leaching of gelatin during swelling only to a small degree (see Table I). Strips aged 39 months at room temperature (runs 10, 5, and 9) lost one quarter more gelatin during a 96-h swelling period than did those aged 2 months (runs 13 and 15). The increased leaching was ascribed to a slight hydrolytic degradation of the strips during storage, where their moisture contents-ranged from 1.7 to 4.0%. The strips of runs 10,5,and 9 had similar properties. The swelling was effected in solutions of different buffers having identical molar concentrations. The amount of gelatin lost determined by colorimetric analysis (run 10) and by weight loss of the strips (runs 5 and 9) are in fair agreement (see Table I), indicating that the nature of the buffer had no significant effect on leaching of gelatin. Gelatin has a broad molecular weight distribution. Presumably, the low molecular weight fractions are dissolved preferentially during swelling. A commercial type B gelatin with a weight-average molecular weight of 95 000 contained 6% of a fraction with a molecular weight of
Table CLeachlng Rate of Oelatln During Sweillng' Amount of Gelatin Dissolvedb % wlw f SDc

Swelling Time, h Run 13 0.5 1.o

0.65 1.0

-t 0.07 f 0.1

Run 15 (2) (2)

1.5

2.0 2.5 4.0 8.0 22.5 26.0 46.0 55.0 70.0 95.0 96.0 102.0

Run 10

Run 5

Run 9

6.1 2 0.3 (3)

7.0 f 0.4 (3)

0.75 f 0.07 (2)

2.0 (1) 2.2 2 0.1 (2) 2.9 f 0.3 (3) 3.9 f 0.5 (3)

0.2 (2)

2.1 f 0.0 (2)

2.0 f 0.1 (2) 2.8 2 0.1 (2)

3.0 f 0.4 (2)

5

4.2 2 0.1 (2) 4.5 '' 0.2 (3)

4.8

5

0.1 (2)

5.1 2 0.6 (2) 5.2 f 0.3 (2) 7.0 f 0.1 (2)

5.6 f 0.2 (2)

*In 0.15 M buffer solution at 20 "C.See Table II for strip properties and buffer. Blank spaces indicate that no measurements were taken at those times. bAt swelling times indicated. CNumbersin parentheses for runs 13, 15, and 10 represent the number of buffer solutions analyzed; for runs 5 and 9, they represent the number of strips dried to determine the loss of gelatin.

became brittle a t moisture contents of <2%.Dried strips that were 0.3 mm thick weighed -0.3 g. After swelling for 4 h, they weighed -2 g and measured -2.3 x 6.0 x 0.18 cm; after 96 h, they weighed -3 g and measured -2.4 x 6.4 x 0.21cm. The results of the swelling experiments are discussed according to the various process variables investigated. In these experiments, one variable was varied at a time while the others were maintained constant. Typical swelling isotherms are shown in Figs. 1-3. Each point represents the average uptake of buffer solutions by 37 strips. Table I1 summarizes the data for 19 such swelling curves, listing solution uptake at 0.5 h, 4.0 h, and a t the longest swelling time employed for each run. All swelling data of Table I1 and Figs. 1-4 were corrected for the small amounts of gelatin loat by leaching at the Corresponding swelling times. All swelling isotherms (Figs. 1-4) consist of a steep initial portion. They begin to level off substantially at 6-10 h and

700

c

0

10

20

30

40

TIME. h

Figure 2-Swelling isotherms of gelatin strips of various thicknesses in 0.15 M ammonium acetate at 20 "C. €ad, point represents the average of 4-5 strips. Key: (A) 0.17-mm thickness (run 11); (A) 0.23-mm thickness (run 12); (0) 0.26-mm thickness (run 13); (0)0.29-mm thickness (run 14).

2ml I00

01

0

'

I

10

'

I

'

20

I

30

"

40

TIME, h

Figure 1-Typical swelling isotherms in 0.15 M ammonium acetate. Key: (A) run 14; (0)run 16; (0)run 16; (A)run 19. 792 / Journal of Pharmaceutical Sciences Vol. 75, No. 6, August 1966

approach the horizontal line corresponding to equilibrium swelling asymptotically. At the longest swelling times routinely employed, 96 h, the uptake was within 4% of that at 168 h. The 96-h value is designated as "near-equilibrium" uptake. Linear Representation of Swelling Data-Equation 1 was obeyed well over the entire swelling period investigated, fmm 0.25 to 96 h; linear correlation coefficientsranged from 0.9923 to 0.9997. Two criteria, in addition to the correlation coefficients, were used to examine the agreement of the data with eq. 1. The first involves using the values of A and B, obtained by linear regression according to eq. 1, to recalculate the extent of swelling a t the various times and to compare it with the observed swelling. Significant deviations were found for all swelling isotherms in the region of maxi-

'

\

9s

<$

oc

Cone.-

15 20 25 15 17

25 25 25 17

17 17 17 17 17 17 17 17 17 17

No.B

1 (3) 2 (3) 3 (3) 4 (3) 5 (3)

6 (3) 7 (3) 8 (3) 9 (3)

lO(3) 1 1 (4) 12 (5) 13(5) 14 (4) 15 (5) 16 (5) 17(5) 18(4) 19 (7)

9 8 5 9

325 2 217% 291% 338'5 3582 329 2 252 2 337 2 330% 326 %

8 9 7 14 4 10

4 3

11

534 2 26 5442 9 560230 322 2 7

366% 340% 336' 321 2

3442 8

Weight. rng 2 SD

Age, months

5.5 5.0 5.0 6.0 39

4.0 3.3 3.4 2.5 1.7

Moisture, %w/w

Residual

0.01 0.02 0.02 0.01

4.5 4.5 4.5 39

0.30 % 0.01 0.172 0.01 0.232 0.01 0.28 2 0.01 0.29 2 0.01 0.29 2 0.01 0.20 2 0.00 0.25 t 0.01 0.272 0.01 0.27 5 0.01 2.0 2.0

11

39 2.0 2.5 2.0 2.0 2.3 0.5

Ammonium Acetate

0.53 2 0.53 2 0.53 2 0.302 1.7 1.8 2.2 3.1 2.7 3.6 3.4 2.8 3.1 2.8

2.4 2.4 3.5 1.7

Triethanolamine Hydrochloride

0.29 2 0.02 0.29 % 0.02 0.30t 0.03 0.28 2 0.02 0.30 2 0.00

Tris Hydrochloride,

Thickness,' rnrn 2 SD

Strip Properties

0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15 0.15

0.10 0.30 0.90 0.15

0.15 0.15 0.15 0.15 0.15

Bur,

20 20 20 20 20 20 20 20 25 20

22 22 22 20

22 22 22 22 20

"C

Swelling

Swelling Conditions

~

~

~~~~

3.122 0.19 4.42 2 0.12 3.52 % 0.12 3.142 0.16 3.01 % 0.06 3.21 t 0.08 4.04 2 0.06 3.022 0.11 3.53 2 0.09 3.09 2 0.06

1.872 0.12 1.78% 0.02 1.58% 0.08 3.37 2 0.10

3.03 5 0.08 3.23 % 0.17 3.232 0.14 3.18 t 0.06 3.31 t 0.11

o,5

6.77t 0.11 6.84% 0.08 6.65% 0.05 6.492 0.03 6.32 2 0.04 6.78 -t 0.07 6.98 2 0.07 6.352 0.08 7.96 2 0.11 6.532 0.11

5.65 % 0.13 5.65 2 0.02 5.502 0.11 7.36 2 0.03

7.23 % 0.08 7.02 % 0.08 7.00 t 0.06 7.23 t 0.08 7.31 % 0.11

4.0h

*

(25) (25) (24) (96)

0.14 (96) 0.10 (96) 0.09 (96) 0.11 (102) 0.02 (72) 9.44 0.07 (95) 9.205 0.09 (96) 8.69 2 0.05 (95) 13.87 2 0.35 (96) 8.69 2 0.02 (72) 9.59 % 8.79 2 8.85 % 9.044 8.754

8.07 2 0.06 8.41 4 0.02 8.83 t 0.04 10.54 2 0.14

10.65 4 0.20 (96)

-

Maximum Timed

Uptake of Swelling Solution, g Solution/g Dry Gelatin

0.101 2 0.007 0.0448 2 0.0010 0.0755 ? 0.0016 0.09532 0.0027 0.103 2 0.003 0.09772 0.0024 0.05872 0.0013 0.0998? 0.0027 0.09472 0.0034 0.0989 2 0.0024

*

0.204 0.006 0.223 ? 0.005 0.260 t 0.009 0.0927? 0.0022

* *

)

0.112 * 0.003 0.100 0.003 0.0926 0.0020 0.104 * 0.003 0.0967* 0.0034

(

A 2 SD,' h . g Gelatin g Solution

0.126 2 0.1352 0.1322 0.1332 0.135 2 0.1252 0.129 % 0.134 2 0.105 2 0.131 2

0.003 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001 0.001

0.128 % 0.003 0.124 2 0.002 0.121 % 0.004 0.114 2 0.001

0.113-C 0.001 0.1192 0.001 0.120% 0.001 0.115 5 0.001 0.114 2 0.002

B 2 SD,' g Gelatin ( m n )

'The strips of run 15 were conditioned for 12 h at 25 "C and 85% relative humidity; all others were conditioned for 72 h. bThe number of strips tested is shown in parentheses. =Basedon five measurements per strip. "The maximum times in hours are given in parentheses. .Intercept in eq.1.'Slope in eq. 1. Both intercept and slope were calculated by linear regression of the initial 4-h swelling data: 1IA representsinitialswelling rate; 1/Brepresentsequilibrium swelling. gMaximumswelling time for runs 1 4 was 4 h.

Gelatin Casting

of Varhbleb In Gelatin Strlp Prepratton and of Swelllng Condhlonr on Uptake, Rae, and Equlllbrlum Swelllng'

Run

Tabb I l - E M

-

55

400

4

200 O

1 0

I

L

I

I

20

I

I

I

I

60

40

I

80

I 100

TIME, h

flgure 3-Swelling isotherms of gelatin strips in 0.15 M ammonium acetate at two temperatures. Each point represents the average of 5-7 strips. Key: (0)25°C (run f8); (0)20°C (run 19).

r

5 I-

400r .-I

300 ,

I00

-

00

2

4

6

8

1020

40

60

80

100

TIME. h

ment of the polymer chains in the crystdlites results in greater lateral interchain attraction and, hence, in a greater elastic force resisting expansion and swelling. In the case of gelatin, the swelling of the amorphous regions is probably complete at the times corresponding to the maximum curvature in Figs. 3 and 4, namely, 6-10 h. At that stage, the swelling of the crystalline domains has probably not proceeded very far, first because it is slower than the swelling of the amorphous domains, and second because it is less extensive, resulting in a smaller maximum uptake. The equilibrium swelling, 1/B = W,, refers mainly to the amorphous regions of the gelatin and is, therefore, smaller than the maximum experimental uptake. The latter approaches the equilibrium swelling for the entire gelatin sample asymptotically. Equation 1 represents a aecond-order process. It hae a single rate constant, which is equal to 11A and represents the initial rate of swelling. The reciprocal values of the experimental intercepts refer overwhelmingly to the swelling rate of the amorphous regions. This process is far more extensive and faster than the swelling of the crystalline regions and is, therefore, much more important for the behavior of gelatin. Consequently, eq. 1 wae applied to the initial 4-h swelling periods which correspond practically entirely to the swelling of the amorphous regions. The A and B values of Table tI were obtained by linear regression of the swelling data of the initial 4-h periods. Each calculation involved 30 to 70 individual data points, measured a t 0.25,0.50,0.75,l.O, 1.5,2.0, 2.5,3.0,3.5,and 4.0 h. Curves I1 and 111of Fig. 4 were calculated with the A and B values from the initial 4-h swelling period and h m the entire 96-hperiod, respectively. Curve I1 matches the experimental isotherm (Curve I) better than Curve 111, except, of course, a t the long swelling periods of 227 h. Figures 5 and 6 reproduce the swelling data of Figs. 1 and 2, respectively, calculated according to eq. 1. The straight lines were obtained by linear regression, using 4-h slopes and intercepts. The agreement between the experimental pointa and the regression lines is satisfactory, confirming the validity of eq. 1. The maximum uptakes at 20°C, shown in the eleventh column of Table II, exceed the 1/B values, which were

Flgure &Comparison of an experimental swelllng isotherm with calculated isotherms. Key: (0) experimental pornts (run 15); (-) Curve I, drawn through experimental points; (- -) Curve 11, CcllcUlated with A and B based on the 4-h swelling period; (-----) Curve 111, calculated with A and B based on the 96-h swelling period.

mum curvature (6-10 h), where the observed uptakes were as much as 13% lower than those calculated; smaller deviations were found in the initial, steep portions of the iaotherms (see Fig. 4). The second criterion involves comparing the reciprocal values of the slope, 1/B = W,, with the maximum experimental uptakes, measured between 72 and 102 h. The greatest difference between the calculated and measured values was 5%; most differences were within 2%. The discrepancies shown by Fig. 4 are ascribed to the structure of gelatin. As early BB the 1930's,X-ray diffraction studies of airdried gelatin" showed it to be partially crystalline. The swelling of semicrystalline polymers in solvents consists of two distinct and simultaneous processes. The penetration of solvent into the amorphous regions is fast and extensive because of their lower density, producing high initial swelling rates and leading to high equilibrium uptakes. The penetration of solvent into the crystalline domains is slower and limited because they are more tightly ordered and have a higher density. The equilibrium swelling of the crystalline domains is considerably leas extensive than that of the amorphous regions because the regular arrange194 /Journal of Pharmaceutical Sciences Vol. 75, No. 8, August 1986

070

0 60

1

c

0 50

E.

040

f

020

0 10

I0

20

30

40

TIME, h

Flgun %Linear regmsion of the swelllng Isotherms of F@. 1 aocording to eq. 1. Key: (A) run 14; (0)run 16; (0)run 18; (A)fun 19.

A

I 060

1

where A is in hours x grams of gelatin per gram of buffer solution and H is in millimeters. Of all the factors affecting the swelling rate, strip thickness was the dominant one. When the intercepts of all 74 strips representing 19 runs are plotted against strip thickness, the points fall on a straight line with surprisingly small deviations (Fig. 71, even though the strips were prepared from casting solutions of various concentrations, conditioned and aged for various times, and swollen in solutions of three different buffers a t various concentrations and temperatures. Linear regression of the data of Fig. 7 gives:

A

= -0.044

(n = 74, r

"

'

0

0

'

20

10

'

30

+ 0.511 H =

(4)

0.9670)

The proportionality between A and H is ascribed to the fact that the initial swelling rate, 1/A, is presumably proportional of the strips, which in turn is to the specific surface area, Sepr inversely proportional to H:

-

SB, =

40

2(W+L) WL

+ -H2

TIME, h

Flgure &Linear regression of the swelling isotherms of Fig. 2 according to eq. 1. Key: (A) 0.77-mm thickness (run 17); (A) 0.23-mm thickness (run 12); (0) 0.28-mm thickness (run 13); (0) 0.29-mm thickness (run 14).

calculated from the initial 4 h swelling data, by an average of 1.34 g of solution per gram of gelatin or 15% (n = 8). This difference represents an estimate of the extent of swelling of the crystalline regions. It is the lower limit because, in the few instances where 168-hvalues are available, they exceed the 96-h value by ~ 4 % . Swelling Constants-The constants A and B of eq. 1 are listed in Table I1 for each swelling experiment. The effects of the various process variables on these two parameters are discussed below. They were considered statistically significant if calculated F ratios (or a t value for a pair of A and B averages) exceeded the critical values at the 5% level. The 95% confidence intervals of the A and B values were then examined to determine which values were different. Gelatin Concentration-The effixt of the gelatin concentration of the casting solutions on swelling rate and equilibrium swelling a t constant strip thickness was evaluated in runs 1 3 . There was no significant difference in swelling rates for casting solutions containing 15, 20, and 25% w/w gelatin. The equilibrium swelling of run 1,with 15%gelatin, was slightly but significantly higher than those of the runs with 20 and 25% gelatin. Gelatin Strip Thickness-In Runs 11-14, only strip thickness and weight were varied. As seen in Table 11, thinner strips swelled faster, taking up more buffer solution per gram of dry gelatin a t 0.5 and 4 h than did thicker ones. This is also shown in Fig. 2, where the top isotherm corresponds to the thinnest strips and the bottom isotherm to the thickest strips. The linear plots in Fig. 6 are parallel; their slopes, B, are not significantly different equilibrium swelling, represented by 1/B, was independent of strip thickness. The intercepts A, whose reciprocal values represent initial swelling rates, are proportional to the initial strip thickness, H. In the range of thickness employed, 0.16-0.53 mm, the relation is:

A

=

-0.029

(n = 18, r

+ 0.453 H =

0.9771)

(3)

All strips were cut to the same width, W, and length, L; only the initial thickness, H, was varied. Thus, the first term of eq. 5 is constant and S , is inversely proportional to H. Even during the initial 4-h swelling period, W and L increased only by -33%. while H increased by 500%: the first term of eq. 5 remained nearly constant throughout that period. Conditioning Time of Gel Strips-The freshly cut strips of gel gelatin were first conditioned a t 25 "Cand 85% relative humidity to remove most of their water gradually, preventing a sudden collapse of the gel structure. Only later were they fully dried over anhydrous CaSOl (Drieritel at room temperature. All gel strips were conditioned for 72 h in the controlled environment, except those of run 15 which were conditioned for only 12 h. Except for the conditioning time, the strips of run 14 resemble those of run 15 quite closely. Therefore, comparison of these two runs provides information on the effect of conditioning time. The shorter.conditioning time resulted in a higher equilibrium swelling but no significant difference in initial swelling rate. The higher equilibrium swelling is because

I

P

7

0 005

010

0 20

030

040

OW

THICKNESS, m m

Figure 7-€ffect of strip thickness on intercept A of eq. 1. Numbers identify runs listed in Table I/. The line was calculated by regression analysis. (Note that the abscissa does not start at zero thickness.) Journal of Pharmaceutical Sciences / 795 Vol. 75, No. 8, August 7986

longer conditioning times afford the gelatin molecules more opportunity to arrange themselves into a crystalline structure, which offers a greater restraint to swelling. Age of Dried Strips-The effect of storage times of the strips over anhydrous CaS04 (Drierite) a t room temperature was evaluated for runs 16, 14, 17, and 10, corresponding to strips aged for 0.5,2.0,11, and 39 months, respectively. Since strip thickness strongly affects the swelling rate and the strips of these runs were not of identical thickness, the observed rates were corrected for differences in thickness with eq. 3. Differences in swelling rates were not significant except for the 11-month old strips (run 171, which showed a 16% reduction in swelling rate compared with that calculated from eq. 3. The slight rate increase of the 39-month old strips (run 10) was not significant. Equilibrium swelling of the 2- and 11-month old strips (runs 14 and 17) was the same within the accuracy of the measurements. It was 4 and 7% higher for the 0.5- and 39month old strips (runs 16 and lo), respectively. The slightly, but significantly, higher equilibrium swelling of the 0.5-month old strips, compared with those aged for 2 and 11 months, is ascribed to further crystallization on storage over anhydrous CaS04 (Drierite) during the initial 2 months, assisted by the 2-3% residual moisture present. This is corroborated by the observation that the density of the 11month old strips was 0.13 g/cm3 higher than that of 0.5month old strips, and by the minimum in swelling rate at age 11months. The higher equilibrium swelling of the 39-month old strips is ascribed to a slight hydrolytic degradation occurring between 11 and 39 months storage (see above), which in turn increased the leaching of gelatin during swelling and may have weakened the restraining network of its chains. Residual Moisture Content-The residual moisture content of strips stored at room temperature over anhydrous CaS04(Drierite)was uniform within a given batch but varied slightly from batch to batch and, hence, from run to run. For instance, the strips used in runs 1 and 4 had 4.0 and 2.5% residual moisture prior to the swelling experiments, after 5.5 and 6.0 months storage, respectively. The small differences in residual moisture had no significant effect on rate and equilibrium swelling. Chemical Nature of Buffer in Swelling Solution-The following three buffers were compared in 0.15 M solutions (at 20 "C)which had pH values of 7.0, using strips from a single batch: triethanolamine hydrochloride (run9), Tris-HC1 (run 51, and ammonium acetate (run 10). The latter produced the lowest uptakes a t 0.5,4, and 96 h, and the lowest equilibrium swelling. There were no significant differences in the initial rate of swelling in the three buffers. The reduced swelling in ammonium acetate may be due to the fact that the acetate ion has a lower lyotropic number than the chloride ion. Therefore, it tends to favor gelation of solutions and resistance to swelling of solid gelatin more strongly than the chloride ion.'6 Buffer Concentration of Swelling SolutionSwelling was compared in 0.10,0.30, and 0.90 M solutions of triethanolamine hydrochloride at 22°C (runs 6-8). The rate of swelling was lowest at the high buffer concentration but not significantly different a t the medium and low concentrations. The uptakes a t 0.5 and 4 h were significantly higher a t the low and medium concentrations than a t the higher concentration, but, surprisingly, the equilibrium swelling values a t the three concentrations were not significantly different.

796 /Journal of Pharmaceutical Sciences Vol. 75, No. 8, August 1986

Temperature of Swelling Solution-Figure 3 shows the swelling isotherms at 20 "C (run 19) and 25 "C (run 18).The initial rates of swelling a t the two temperatures are not significantly different. The initial swelling rate of gelatin that was not crosslinked was reported to be practically constant in the temperature range of 11-25 "C.6 The swelling isotherm a t 20 "C is flatter than the one a t 25 "C, indicating lower equilibrium swelling and lower uptakes a t intemediate swelling times. Of all the variables examined, the temperature increase from 20 to 25 "C produced the greatest increase in equilibrium swelling. This effect is caused by a rapid approach to the gel melting point as the temperature is raised through this 5 "-interval. The melting points at pH 7.0 were 34.5 "C for a 17% (w/w) gelatin gel and 36 "C for one of 25%. Moreover, strips swelled a t 25 "C were softer and weaker than strips swelled a t 20 "C, indicating less resistance to swelling. Crystallization of Gelatin-The ordering of gelatin molecules leading to crystalline domains began in solution. It progressed extensively during conditioning at 85% relative humidity and 25 "C, where the cast films set to gels and lost most of their water, and continued slowly during storage at room temperature over anhydrous CaS04 (Drierite). Longer conditioning times of the wet strips led to slower drying, providing the chains with greater opportunity to crystallize. The greater crystallinity on conditioning was confirmed by X-ray diEractionl6 and is corioborated by reductions in equilibrium swelling. During storage over anhydrous CaS04, the moisture content of the strips was reduced to 2-4%. This small amount of water, acting as a plasticizer, provided sufficientmobility for continued slow ordering, because longer storage times further increased the density of the stored strips and reduced the swelling rate.

References and Notes 1. Yannas, I. V. J . Macromol. Sci., Rev. Macromol. Chem. 1972,C7, 49-1 04. 2. Ames, W. M. J . Sci. FoodA ric 1952,3,454-463. 3. Simms, W. M.; Blake, J. N. f;afure 1960,187,998. 4. Jopling, D.W. Research (London)1953,6,27%29S. 5. Jo ling, D.W. J . Appl. Chem. 1956,6,79-84. 6. Ligicky, A.; Bermane, D. in "Photogr.Gelatin, Proc. Sym 2nd 1970";Cox, R. J., Ed.; Academic Press: London, 1972 pp &-48. 7. Lloyd, D. J.; Pleass, W. B. Biochem. J . 1927,21,1352-1367. 8. Sterman, M. D.;Faust, M. A.; Genova, D. J.; Curme, H. G.; Johnson, M. F. in "Photo Gelatin, Proc. Symp., 2nd 1970"; Cox, R. J., Ed.; Academic L s e : London, 1972;pp 113-120. 9. Robinson, I. D.Photogr. Sci. Eng. 1964,8,220-224. 10. Claes, F. H.; Boulonne, A.; Beels, R. Photogr. Sci. Eng. 1978,22, 28-37. 11. Ofner 111, C. M.; Schott, H. J . Pharm. Sci. 1985,74, 1317-1321. 12. Schacterle, G.R.;Pollack, R. L. A d . Eiochern. 1973,51,664655. 13. Williams, J. W.; Saundere, W. M.; Cicirelli, J. S. J . Phys. Chem. 1954.58. 774-782. 14. Hehana, P. H. in "Colloid Science", vol. 11; Kruyt, H. R., Ed.; Elsevier: New York, 1949;chap 12. 15. McBain, J. W. "Colloid Science"; D. C. Heath and Co.: Boston. 1950;chap 9. 16. Bradbury, E.; Martin, C. Proc. Roy. Soc. Lond. 1952,A214,183r 192.

Acknowledgments Financial su port by Max Schwartz, R.Ph., is gratefully acknowledged. The geratin samples were a gift of Smith Kline Beckman Corporation. Adapted in pae t o m a thesis for submission by C. M. Ofner I11 to Temp e University in partial fulfillment of the Doctor of Philosophy degree requirements.