Printer Friendly

A strategy for determining leachables in liquid drug products.

The most frequent application of HPLC analytical methods for the pharmaceutical industry has been to monitor the chemical stability of the ingredients in various drug products. The second most frequent application has been the detection and quantitative determination of process impurities. The process impurities included leachables resulting from product contact with either the packaging materials or the surfaces of the manufacturing equipment. The development of HPLC methods for determining stability and process impurities many times have occurred independently, resulting in an inefficient use of resources (time, personnel and instrumentation) to bring new drug products to market. This inefficient utilization of resources can be overcome by using a combinatorial strategy in developing HPLC methods for pharmaceutical products. The combinatorial strategy is simply the coordinated expansion of the chromatographic and the detection ranges for a given HPLC procedure to cover analytes that are inclusive of formulation ingredients and potential extractable compounds.

This strategy can be executed in two coordinated steps. The first would involve starting with a method developed for determining drug product impurities, then identifying operating parameters that can expand the chromatographic range to provide additional separation capability for extractable compounds. The determination of an expanded chromatographic range needed to accommodate the different analytes would be based upon data derived from both experiments and published literature. In the second step, an expansion of the detection range would be required due to the chemical diversity of potential compounds that could migrate into the formulation as a result of drug product exposure to multiple contact surfaces during manufacturing and packaging. Identifying and developing HPLC methods capable of meeting these chromatographic and analytic objectives is a challenge. Strategies will be presented in this article for addressing these challenges of developing HPLC methods capable of providing analyses for formulation and process impurities.

Results and discussion

Fundamentally, liquid chromatography is a process for separating compounds in a mixture, which takes place at a solid-liquid interphase. The separation occurs by matching the chemical properties of the given compounds with those of either the liquid or solid (stationary) phases, resulting in individual compounds preferentially partitioning with one of the two phases (like to like). The chemical properties of the two phases also provide an analytical tool for characterizing the particular compounds based upon their preferred partitioning or retention with one of the phases. The thermodynamic driving force for the preferred retention can be described in general terms by the hydrophilic or the hydrophobic properties of the solid-liquid phases and the compound of interest. This same driving force is also in play when it comes to describing the process of leachable compounds migrating or partitioning into a liquid drug product. For example, the potential for a plasticizing agent to migrate from the surface of a polymer into a liquid drug product can be estimated (ref. 1) based upon the hydrophilic or hydrophobic properties of the compound and interphase matrices. Thermodynamically, the partitioning driving force is commonly described in terms of a Log P value.

Log P values historically have been determined by measuring the solubility ratio value for a given compound at a liquid-liquid interface. The most commonly used liquid-liquid interface is one created by two immiscible solvents, e.g., octanol and water. The calculated Log P ratio value for a compound that prefers to partition into the organic octanol phase is a positive number (hydrophobic), while a negative number indicates a preference for the aqueous phase (hydrophilic). If however, the compound of interest also has an accessible pKa value (i.e., ionization potential) in the water phase, then pH can affect the partitioning ratio value. This effect would translate into a series of partitioning or distribution (Log D) values. So instead of a single Log P value, there is a series of Log D values as controlled by the pH of the aqueous phase and the pKa value for the compound. Consequently, the partitioning properties can shift for a given compound from hydrophilic (negative Log D value) to hydrophobic (positive Log D value) depending on solution pH.

Log P and Log D values for a compound can also be determined at a solid-liquid interphase using either normal or reverse phase liquid chromatography (ref. 2). The chromatographically determined partitioning values would be based upon a comparison of retention times for a standard, with known Log P or Log D value, and that for a sample. The operating parameters of flow rate, mobile phase composition (e.g., pH) and column temperature must be controlled during the analysis to minimize variances between the standard and sample determinations.

The relationship between Log D values and HPLC retention times can also be applied in designing a single chromatographic method to monitor impurities related to the formulation ingredients and leachable compounds. This process would begin by first focusing on the possible leachable by compiling a list of process related impurities that have been reported in the literature. Several review articles (refs. 3-7) and reference books (refs. 8-10) contain lists of reported extractable compounds from polymeric materials which are commonly used in pharmaceutical manufacturing and packaging. The next step would be to determine the Log D values for the potential leachable compounds in the compiled list, along with the formulation ingredients. The values would be determined either experimentally through a search of the literature or calculated using computer modeling programs for chemical structural/property analysis (e.g., ACD/lab physical properties). The generated data for the listed compounds would then be used to set a chromatographic range needed to resolve the different hydrophilic to hydrophobic compounds, as well as to indicate the potential for analyte co-elution (i.e., compounds having similar Log D values).

[FIGURE 1 OMITTED]

An example of the process for evaluating the Log D values relative to chromatographic performance is shown in figure 1, where log D values for a set of 36 potential leachables compounds are plotted for three pH conditions (acidic, neutral and alkaline). Analysis of these histograms, as shown in table 1, indicate that for an acidic formulation (i.e., pH 4 range) the chromatographic method would need to be able to resolve predominately hydrophobic compounds in this example, compared to formulations that are more alkaline (pH >7-10), where the method would need to separate both hydrophilic and hydrophobic compounds. Additionally, analysis of the histograms also suggests that if the log D values for the formulation ingredients are between 3-4, there is the potential for co-elution with a possible leachable compound.

The extractable compounds reported in the literature could be used as well as calibrators to experimentally assess the capability of a given HPLC method to monitor the range of potential process impurities. The selection of compounds to use as calibrators, from the compiled list of reported extractables, would be dependent upon their log D properties. For example, given a formulation containing a pH 6 buffer and an active ingredient with a log D value of 2, the appropriate set of compounds to use as calibrators would include those with log D values ranging from <1 to 6, especially one with a log D value of two in the range of pH 6. A list of possible compounds that would be suitable for this example is shown in table 2. Once the series of compounds has been selected to use as calibrators, the next step is to experimentally determine their retention times for the particular HPLC method, as shown in figure 2. The relationship between the retention time data and log D values can then be used to determine the chromatographic range and resolution of the given method. This relationship can be assessed by plotting the data, as shown in figure 3, then fitting it to an equation. In this example, the plotted data are well correlated to a linear equation with an [R.sup.2] value >0.9. (Note: Linear regression equations may not always be the best fit for the plotted retention time vs. log D data. Changes in chromatographic operating parameters such as flow rate, shape of mobile phase gradient, column temperature, etc., can result in a better fit of the data to a non-linear polynomial equation.)

[FIGURE 2 OMITTED]

Conventionally chromatographic resolution ([R.sub.s]) is calculated as a function of differences in retention between two analyte peaks ([t.sub.1] and [t.sub.2]) and the average width of the peaks ([W.sub.avg]) all in terms of time measurements as shown in equation 1. The [W.sub.avg] term in equation 1 can be defined in terms of time and theoretical plate number (N), as shown in equation 2.

[R.sub.s] = ([t.sub.2] - [t.sub.1])/[w.sub.avg] (1)

[w.sub.avg] = ([w.sub.1] + [w.sub.2])/2 = (4[t.sub.1]/[square root of (N)] + 4([t.sub.2]/[square root of (N)])/2 = 2([t.sub.1] + [t.sub.2])/[square root of (N)] (2)

Substituting the [t.sub.1], [t.sub.2] and N terms for [W.sub.avg] in equation 1, the [R.sub.s] equation can then be calculated in terms of time and theoretical plate number as shown in equation 3.

[R.sub.s] = ([t.sub.2] - [t.sub.1])[square root of (N)]/2([t.sub.2] + [t.sub.1]) (3)

The retention time values for a compound can also be described in terms of log D, as demonstrated in figure 3, where in the presented case the relationship between the two terms can be defined by a linear regression equation. As shown in equation 4, retention time ([t.sub.i]) value is directly proportional to the slope of the regression line (s), the log D value ([D.sub.i]) and the intercept (a) of the regression line.

[t.sub.i] = sLog[D.sub.i] + a (4)

Combining the terms in equation 4 with equation 3, then collecting and simplifying the terms in the equation, results in an expression, as seen in equation 5, that redefines the [R.sub.s] factor as a function of Log D values, theoretical plate number (N) and the intercept for the regression equation.

[R.sub.s] = (Log([D.sub.2]/[D.sub.1])[square root of (N)]/2(Log([D.sub.2]/[D.sub.1])+2a) (5)

Defining the [R.sub.s] factor in terms of Log D provides a way to determine the number of compounds with different partitioning values that can be monitored for a given chromatographic method. The probability of resolving k number of randomly spaced chromatographic peaks between the two reference peaks, as shown in equation 6, for a single detection system, can be calculated in terms of the [R.sub.s] factor that was determined in equation 5. (Note: Where [R.sub.s] values are rounded to the nearest integer.)

% probability ([P.sub.1]) = 100[[R.sub.s]!/[R.sub.s.sup.k][R.sub.s] - k)!] (6)

Applying then equation 5 to the chromatogram shown in figure 2 results in a calculated [R.sub.s] factor equal to 21 for the separation between the DCBA and BHT peaks. Using the Rs value of 21 in equation 6 results in a calculated probability value of 70% for the resolution of four peaks between these two reference HPLC peaks. If the number of peaks to be resolved increases from four to six, the calculated percent probability value drops off exponentially to a value less than 50%, as shown in table 3. If the [R.sub.s] factor were to increase from 21 to 50, by expanding the retention time window between the two reference peaks (examples would be reducing flow rate or change gradient profile of the mobile phase), the calculated probability for resolving six peaks would correspondingly increase from 50% to 70%. However, the downside in expanding the retention time window for a given chromatogram is that it can result in broadening the HPLC peak width by reducing the theoretical plate heights of the column (ref. 1l). Broadening the peak width, especially for the later eluting peaks, would result in a reduction in analyte sensitivity by having lower peak heights and increase the potential for peak-to-peak overlay.

[FIGURE 3 OMITTED]

Another approach that can be taken to improve the probability of detecting more peaks without altering the [R.sub.s] factor is to replace the single detector LC system with a dual detector system. In the dual system, each detector would operate independently of the other detector. The addition of a second independent detector results in a modification of the initial probability equation for PI, to a two-dimensional probability equation for [P.sub.2], as shown in equation 7.

% probability ([P.sub.2]) = 1 - [[1 - [P.sub.1]].sup.2] (7)

An example of a dual independent detector system coupled to an HPLC instrument would be the combined use of UV/visible and mass spectral detectors. In the dual detector example, the 70% probability is extended to the nine peak resolution level, which is more than twice the number of peaks achieved by a corresponding single detection system, as shown in figure 4. These results demonstrate that the use of two independent detectors provides a similar increase in analytical resolution compared to that predicted for an expanding retention time window, but without the chromatographic band broading effect.

If the use of two detectors is better than one detector, can further improvements in resolution be achieved by using three independent detectors? The probability calculation for using three or more independent detectors ([P.sub.3] to [P.sub.n]) is similar in form to that for the dual detectors, as shown in equation 8.

% probability ([P.sub.3]) = 1 - [[1 - [P.sub.1]].sup.3]

to

% probability ([P.sub.n]) = 1 - [[1 - [P.sub.1]].sup.n] (8)

The calculated probability for resolving k number of peaks using three or more detectors is not significantly different than using a dual detection system as shown in figure 4 (Rs 21 (2) vs. Rs 21(3)). In short, no major improvement in peak resolution is achieved by expanding the number of detectors from two to three.

In summary, the relationship between the retention time value for a given compound and its Log D property provides a means for measuring both the analytical range and the resolution capability of a given chromatographic method, in which analyte separation is primarily dependent upon the hydrophilic/hydrophobic interaction between the mobile and stationary phases of the HPLC column. In particular, the chromatographic range determination is based upon the measurement of the retention of the most negative to the most positive Log D compound for a reverse phase column and inversely, positive to negative, for a normal phase column. Chromatographic resolution, on the other hand, can be determined by a measure of the difference in retention time values for two analytes having different Log D values. These differences in retention times in proportion to the variation in Log D values for three or more analyte peaks provide a calibration of the resolution for the particular chromatographic method. Chromatographic resolution can further be evaluated in terms of the number of independent detectors used, which adds another dimension to the assessment. Applying all these concepts to the development of chromatographic methods for pharmaceutical products can result in analytical procedures with expanded capabilities and efficiencies.

[FIGURE 4 OMITTED]

Besides contributing to improving the development of chromatographic methods, the strategy of using partitioning data and independent dual detection systems can also provide efficiencies in identifying unknown impurity compounds in liquid drug products. Examples will be shown of how the analytical tools of partition [Log D/retention time] calibration and dual spectral detection [e.g., UV and MS] have been used in a synergistic approach to identify unknown leachable compounds that were produced as a result of a pharmaceutical drug product being exposed to two different rubber surfaces. One example is the exposure of a formulation to a Viton rubber o-ring (manufacturing equipment) and the other is an exposure to a halide butyl rubber stopper (primary packing).

Case study 1

As part of any selection process for qualifying a particular piece of equipment for use in the manufacturing process for a new drug substance or product, sample materials representing the surfaces of the identified equipment would be tested for compatibility and extractables. The extractable test would involve the incubation of the material in solvents or solutions, which serve as surrogates for the reagents used in the manufacturing procedure, under stress conditions. After incubation, the surrogate reagents can be analyzed for potential leachables that could occur during the manufacturing process. In one case, an extractable compound was detected in an aqueous buffered solution that was exposed to a Viton o-ring. The o-ring served as a seal between two stainless steel transfer lines. The UV spectral scan of the extractable compound is shown in figure 5. It closely corresponds to the spectral pattern for a class of substituted t-butyl phenol compounds. These compounds are frequently used as antioxidants to protect polymeric materials. In general, this group of compounds has a characteristic UV absorption pattern with a primary [[lambda].sub.max] peak between 222 nm and 224 rim, and a secondary [[lambda].sub.max] peak between 275 nm and 280 nm. Two examples of this group of antioxidant reagents are BHT (butylated hydroxytoluene) and Irganox 1076 (octadecyl-di-t-butyl-hydroxyhydrocinnamate). The molecular structures, formula weights and Log D values for these t-butyl phenolic antioxidants are shown in figure 6.

[FIGURE 5 OMITTED]

[FIGURE 6 OMITTED]

[FIGURE 7 OMITTED]

Information provided by the manufacturer of the "O-ring" seal did suggest that the bulk polymer used to fabricate the seals could contain BHT and/or Irganox 1076. The rationale for using these two antioxidant reagents in the polymer is that the respective positive Log D values for BHT at 5.3 and for Irganox 1076 at 13.9 are predictive that they have a very low probability of partition into an aqueous solution from a hydrophobic polymer matrix.

The HPLC retention time for the extractable peak indicates that it has a Log D value estimated to be 3.8, when compared to the retention times of compounds with known Log D values shown in table 2. While a positive Log D value of 3.8 indicates that the extractable is also a hydrophobic molecule, it is significantly less hydrophobic than either BHT or Irganox 1076. The mass spectral data for the extractable compound further distinguishes it from these two antioxidant reagents. The reported mass selective ion peak for the extractable compound is 279 m/z, when operating in positive ionization mode (M+H). Compared to the formula weight for BHT at 220.35, the formula weight for the extractable compound at 278 is significantly larger by 58 mass units. This difference in mass units does not readily suggest a type of functional group that could be attached to the BHT molecule. Compared to the formula weight of lrganox 1076 at 530.86, it is 252 mass units lighter. This difference in formula weight, however, would correspond to the loss of a C18 aliphatic chain on the esterified Irganox 1076 molecule, which would then produce a t-butyl hydroxybenzenepropanoic acid species, as shown in figure 7.

The calculated Log D (calculation via Advanced Chemistry Development (ACD/Labs) software version 10.0 (2007)) value for the proposed aromatic carboxylic acid compound is 3.88 in the pH range of 4.5 to 5.0, which closely corresponds to the Log D value for the unknown compound based upon an analysis of retention time data, as shown in table 4.

A tentative identification of the unknown extractable compound was made based upon the match of the molecular weight, consistence of the UV spectral scan and the partitioning ratio value. Confirmation of the identified extractable compound was achieved by a direct comparison of the HPLC, UV and mass spectral properties to an authentic sample.

Case study 2

The process for determining the suitability of a particular material for packaging a new drug product follows a similar approach as that used for the selection of the manufacturing equipment. The selection process for the packaging material begins by filling containers that are to be tested with solutions, which serve as surrogates for the drug product, then sealing the containers with closures that are also to be tested. A portion of the sealed containers is stored in an upright position to test the effects of the container, while another portion is stored inverted to assess the effects of the closures. Both sets of test samples are held under environmental stress conditions. The liquid stored in these containers and closures is analyzed for extractables at specified intervals of time. An example of a chromatogram for an aqueous solution exposed to a halide butyl elastomeric stopper is shown in figure 8, where both UV and MS detectable impurities were observed.

The retention time values for the five impurity peaks suggest that the Log D values for these compounds range from 1 to 5, based upon comparison of retention times to the reference compounds listed in table 2. The estimated Log D values, along with the UV and the MS spectral data for these five impurities, are summarized in table 5.

As shown in figure 8, three of the five impurities were detected in both the UV and the MS chromatograms, which provide a dual analytical approach for identifying the leachable compounds. The two early eluting peaks, however, were not detected in both chromatograms. Impurity peak labeled #1 was detected only in the UV chromatogram, while impurity peak #5 was detected in only the MS chromatogram. The ability then to identify these two compounds appears to be limited to a single analytical approach.

In the case of impurity #5, at a retention time of 5.7 min., the MS spectral signal provides information about the molecular ion, while the absence of a UV signal suggests that the compound may be aliphatic. However, when the amount of sample injected on to the column was increased, a UV absorption peak at the 5.7 min. retention time was detected. The UV spectral profile for this peak did contain two distinctive [[lambda].sub.max] values, one at 220 nm and the other at 275 nm, indicating the presence of an aromatic or a conjugated moiety (ref. 12). The initial absence of a UV signal response is attributed to sensitivity of the analytical method and not to chemical structure.

[FIGURE 8 OMITTED]

The UV spectral scan for impurity peak # 1, at a retention time of 9.7 min., likewise produced two distinctive [[lambda].sub.max] values at 245 nm and 286 nm. These two [[lambda].sub.max] values distinguish it from impurity peak #5, but also indicate that it contains an aromatic or conjugated moiety.

Further determination as to the identity of impurity peak #5, based upon the available HPLC data, is problematic due to the absence of an MS peak. No MS chromatographic peak could be detected at the 9.7 min. retention time, whether operating in either positive or negative ionization mode, as well as by increasing the amount of sample injected onto the column. An alternate approach to determine the molecular mass of the impurity is the use of GC/MS analysis. The selection of the GC/MS technique is based upon the assumption that the compound of interest is a volatile organic species. The aqueous sample was prepared for GC analysis using a solid phase micro extraction (SPME) technique to transfer the compounds of interest to a more suitable volatile matrix. A representative GC/MS ehromatogram of the impurities linked to the stopper is shown in figure 9. Three of the five impurities detected in the HPLC chromatograms, as shown in figure 8, were identified in the GC chromatograms. The identification was based upon a comparison of the MS scans of the peaks in the blank matrix to the peaks in the stopper extract samples. The molecular ion mass for impurity peak #1 was then determined by GC/MS to be 160 m/z.

[FIGURE 9 OMITTED]

The MS determined molecular ion values did also provide information about the atomic composition of these compounds using the "nitrogen exclusion rule" (ref. 13). For example, in the case where the molecular ion value (m/z [M]) is an even number, then the number of nitrogen atoms in the given analyte is expected to be equal to zero or some even number of nitrogen atoms. When the molecular ion value is an odd number, then the number of nitrogen atoms is expected to be equal to at least one or some odd number of atoms. Applying the rule to the MS data for the five impurities, impurity peak #1 would be expected to contain either no or an even number of nitrogen atoms, since it has an even number molecular ion (m/z[M]) value of 160. The four other impurities would be expected to contain at least one or an odd number of nitrogen atoms since the molecular ion (m/z[M]) values for these compounds have odd numbers of 135, 169, 209 and 197.

The combined UV and MS spectral data, along with the Log D partitioning values, provide then a three parameter reference for tentative identification of the five compounds. Comparing the compiled data for aqueous extractable compounds with the list of reported extractables from rubber stoppers reported in the literature, as shown in Table 6, the following tentative identifications of the five compounds are:

* Impurity 1: Benzaldehyde:

1. Exact match to UV and MS spectral profiles.

2. Does not contain a nitrogen atom (nitrogen exclusion rule for even number MI).

3. Log P value is 1.64 compared to estimated value of < 2.1.

* Impurity 2: Benzothiazole:

1. Exact match to UV and MS spectral profiles.

2. Contains at least one nitrogen atom (nitrogen exclusion rule for odd number MI).

3. Log D value (at pH >5) is 2.01 compared to estimated value of 2.1.

* Impurity 3: Diphenylamine:

1. Exact match to UV and MS spectral profiles.

2. Contains at least one nitrogen atom (nitrogen exclusion rule for odd number MI).

3. Log D value (at pH >5) is 2.97 compared to estimated value of<3.9.

* Impurity 4: Dimethyl dihydroacridine:

1. Similar UV and MS spectral profiles to acridine.

2. Contains at least one nitrogen atom (nitrogen exclusion rule for odd number MI).

3. Log D value (at pH >5) is 4.19 compared to estimated value of<5.3.

* Impurity 5: Dibenzylamine:

1. Exact match to UV and MS spectral profiles.

2. Contains at least one nitrogen atom (nitrogen exclusion rule for odd number MI).

3. Log D value (at pH >5) is 0.32 compared to estimated value of 0.3.

[FIGURE 10 OMITTED]

The chemical structures of these compounds are shown in figure 10. Confirmation of the identity of these impurities was verified by spiking the aqueous solution with the five listed compounds, which produced chromatograms that matched those in figure 8.

Conclusions

Improvements in the efficiency of conducting pharmaceutical development programs for liquid formulations can be achieved by expanding the analytical scope of the chromatographic methods that are initially designed to monitor impurities related to the formulation ingredients. The analytical focus for monitoring impurities thus needs to be inclusive not only of the formulation ingredients, but also for potential leachable compounds which result from drug product exposure to the surfaces of the processing equipment and/or primary packaging materials. The Log D (or Log P) partitioning values for the impurities due to ingredients and process can provide guidance for developing chromatographic methods that are capable of this expanded analytical role.

In addition to contributing to the design of chromatographic methods, Log D (Log P) partitioning data can also contribute to the identification of unknown impurities, as demonstrated in the presented case studies. The combination of partitioning data with data obtained from two independent spectral detection systems (e.g., UV and MS) provides a three parameter approach for identifying unknown compounds using readily available commercial instrumentation.

Adopting then a broader analytical scope during the initial stages of the formulation selection program and incorporating partitioning values as an investigative tool for identifying unknown impurities can provide early detection of leachables, avoiding later program delays and supporting product safety claims.

References

(1.) D.R. Jenke, "Linking extractables and leachables in container/closure applications, " J. Pharm. Science and Tech. 2005, 59, 265-282.

(2.) R.M. Carlson, R.E. Carlson and H.L. Kopperman, "Determination of partition coefficients by liquid chromatography, "J. Chromatography 1975, 107, 219-223.

(3.) D.R. Jenke, "Extractable/leachable substances from plastic materials used as pharmaceutical product containers/devices, "J. Pharm. Science and Tech. 2002, 56, 332-371.

(4.) D.R. Jenke, "Extractable substances from plastic materials used in solution contact applications." An updated review, "J. Pharm. Science and Tech. 2006, 60, 191-207.

(5.) D.R. Jenke, D. Zietlow, et al., "Accumulation of organic compounds leached from plastic materials used in biopharmaceutical process containers, " J. Pharm. Science and Tech. 2007, 61, 286-302.

(6.) J.F. Castner, N. Williams and M. Bresnick, "Leachables found in parenteral drug products," Am. Pharmaceutical Rev. 2004, 7, 70-75.

(7.) J.F. Castner, J. Anderson and P. Benites, "Strategy for development and characterization of HPLC methods to investigate extractables and leachables (part II), " Am. Pharmaceutical Rev. 2007, 10, 10-18.

(8.) Plastic and Rubber Additives, Edition 2005, Synapse Information Resources, Endicott, NY 2005 (ISBN 1-890595-75).

(9.) K.C. Waterman, R.C. Adami and J. Hong, Impurities in Drug Products (Chapter 4), S. Ahuja and K.M. Alsante , Eds., Handbook of Isolation and Characterization of Impurities I Pharmaceuticals Volume 4, Academic Press, NY, 2003, (ISBN 012044982X, 9780120449828).

(10.) Zweifel, Hans, et al., Plastics Additives Handbook 5th Edition, Hanser Gardener Publications, Cincinnati, OH, 2001, (ISBN 3-446-19579-3).

(11.) L.R. Snyder and J.J. Kirkland, Introduction to Modern Liquid Chromatography 2nd Edition, Wiley-Interscience, N.Y., 1979 (ISBN 0-471-03822-9).

(12.) E. Pretsch, B. Buhlmann and C. Affolter; Structure Determination of Organic Compounds, Springer-Verlag, NY, 2000 (ISBN 3-540-67815-8).

(13.) F.W. McLafferty and F. Turecek, Interpretation of Mass Spectra 4th Edition, University Science, Sausalito, CA, 1993 (ISBN 0-35702-25-3).

by Jim Castner, Michael Bresnick and Pedro Benites *, Lantheus Medical Imaging
Table 1--range of log D values for potential
teachable compounds as a function of pH

 Mean Median Range *

[Log D.sub.PH 4] 4.71 3.64 0.6 to 5.7
[Log D.sub.PH 7] 4.23 2.94 -0.3 to 5.4
[Log D.sub.pH 10] 4.01 2.94 -0.7 to 5.0

* 10% to 80% quartile

Table 2--reference extractables compounds

Compounds CAS # Log D Source/application

2,4-dichlorbenzoic 50-84-0 -0.34 Silicone tubing
 acid [DCBA] to 2.7 * peroxide catalyst
2-(2-butoxyethoxy) 124-17-4 1.79 ** Silicone tubing
 ethyl acetate [BEEA] Pt catalyst binder
1,3,5-triazine-2,4-di 29529-99-5 0.3 to Crosslinking
 thione [DBAT] 2.11 * agent
4-pentyl phenol 87-26-3 3.88 ** Lubricant,
 stabilizer,
 antioxidant
Phenol,2,6-bis 489-01-0 4.69 ** Antioxidant
 (1,1-dimethylethyl)
 -4-methoxy [BHA]
Dibutyl phthalate 84-74-2 4.83 ** Plasticizer
 [DBP]
Butylated 128-37-0 5.32 ** Antioxidant
 hydroxytoluene [BHT]
Dioctyl phthalate 117-81-7 8.7 ** Plasticizer
 [DOP]

* Log D value over a pH range of 4.0 to 7.0

** Non-ionizable compounds over pH range 2 to 12
(Log D = Log P)

Table 3--calculated probabilities for
resolving chromatographic peaks as a
function of [R.sub.S] factor and number of detectors

 [R.sub.S] [R.sub.S] [R.sub.S] [R.sub.S]
Peaks 21(1) * 21(2) ** 21(3) *** 50(1) ****

 2 95.2 99.8 100.0 98.0
 3 86.2 98.1 99.7 94.1
 4 73.9 93.2 98.2 88.4
 5 59.8 83.8 93.5 81.4
 6 45.6 70.4 83.9 73.2
 7 32.5 54.5 69.3 64.4
 8 21.7 38.7 52.0 55.4
 9 13.4 25.1 35.1 46.5
10 7.7 14.8 21.3 38.2
11 4.0 7.9 11.6 30.5

* [R.sub.S] = 21 for 1 detector; ** [R.sub.S] = 21 for 2 detectors;
*** [R.sub.S] = 21 for 3 detectors; and **** [R.sub.S] = 50 for
1 detector

Table 4--determination of Log D value for
unknown extractable compound based upon
comparison to reference compounds

Compound Retention time (min.) Log D *

DCBA 13.2 2.7
(2,4-dichlorbenzoic acid)
Unknown 14.8 3.8 **
4-pentyl phenol 15.1 3.9
BHT 23.4 5.3
Irganox 1076 >60 min. 13.9

* Log D at pH conditions between 4.5 to 5.0.
** Calculated from ACD/Lab Physical Chem Program
Version 10.0

Table 5--summary of the HPLC data for the
impurities detected in aqueous sample

 Retention Log D * UV peaks MSI ***
Impurity time (min.) (pH 5.8) (nm) (m/z [M+H])

1 9.67 -2.1 245, 286 NI ***
2 10.76 ~2.1 225, 252, 285 136
3 22.02 <3.9 285 170
4 25.13 <5.3 237, 282 210
5 5.72 0.3 220, 275 198

* Estimated by comparison to retention time vs. log D values
in table 1

** MSI = mass selective ion (positive mode [M+1])

*** NI = not ionizable both in positive and negative mode
using HPLC/electrospray technique

Table 6--list of aqueous extracted
compounds reported in the literature
along with molecular formula and mass
as well as calculated Log D values

 Mol. Log D
Compound * Formula weight (ph>5) **

Benzaldehyde [C.sub.7][H.sub.6]O 106.12 1.64
Benzothiozole [C.sub.7][H.sub.5]NS 135.19 2.01
Cyclohexanone [C.sub.6][H.sub.10]O 98.14 0.76
Diphenylamine [C.sub.12][H.sub.11]N 169.22 2.97
9,10-dihydro-9, [C.sub.15][H.sub.15]N 209.29 4.19
 9-dimethylacridine
2,6-diterbutyl- [C.sub.15][H.sub.24]O 220.35 5.32
 methylphenol
Dibenzylamine [C.sub.14][H.sub.15]N 197.27 0.32
Ethylbenzene [C.sub.8][H.sub.10] 106.17 3.21
2-butoxyethanol [C.sub.6][H.sub.14][0.sub.2] 118.17 0.80
Tetrachloroethane [C.sub.2][H.sub.2][Cl.sub.2] 167.85 2.17
O-xylene [C.sub.8][H.sub.10] 106.17 3.14
N,N-dibutylacetamide [C.sub.10][H.sub.21]NO 171.28 2.44
Dibutylformamide [C.sub.9][H.sub.19]NO 157.25 2.18
Acetophenone [C.sub.8][H.sub.8]O 120.15 1.66
2-phenyl-2-propanol [C.sub.9][H.sub.12]O 136.19 1.73
2,2,5,5-tetramethyl- [C.sub.8][H.sub.16]O 128.21 2.39
 tetrahydrofuran
Tetrachloroethylene [C.sub.2][Cl.sub.4] 165.83 2.95

* See reference 3

** Values based upon Advanced Development (ACD/Labs)
Software Version 10.0 (2007)
COPYRIGHT 2009 Lippincott & Peto, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Castner, Jim; Bresnick, Michael; Benites, Pedro
Publication:Rubber World
Date:Jun 1, 2009
Words:5805
Previous Article:Facile synthesis of functionalized nano-silica: effect of microstructure on the physical properties of LSR.
Next Article:Barcelona hosts masterbatch 2009.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters