TITLE
SILICA-BASED, ENDCAPPED CHROMATOGRAPHIC PACKING HAVING IMPROVED STABILITY UNDER HIGH PH CONDITIONS
FIELD OF THE INVENTION
The invention is directed to endcapped chromatographic media having improved stability under high pH conditions. In particular, the invention is directed to a process for improving the stability at pH of greater than 7 of silica-based chromatographic media which have been silanized with organosilyl groups.
BACKGROUND OF THE INVENTION
The most popular type of high performance liquid chromatography utilized today is reversed-phase chromatography. Historically this refers to a technique in which the stationary phase is nonpolar and the mobile phase is polar in nature, usually being aqueous. To obtain a nonpolar(or lipophyllic) surface on a support one must attach a nonpolar material to the surface of a rigid support, usually porous silica. The availability of high performance chromatographic packings using nonpolar ligands which are chemically bonded on to a silica support (rather than just physically adsorbed) dates from the late 1960's and early '70 's. These ideas and techniques appear frequently in the literature of this period. An actual chemical bond between the ligand and the silica surface greatly enhanced the stability and reproducibility of the packings.
Chromatographic packings which are used to separate biological samples such as proteins and peptides are routinely subjected to cleaning solutions which can be as high as pH 12. This is done to remove pyrogens from the column and remove strongly absorbed protein from the surface. On an analytical scale this is tolerable since the cost of columns is not large. However, on a process scale, the importance of stability is increased since large amounts of packing are being used and the need to unpack and repack a column involves large amounts of labor and mobile phase. Not only does one have to worry about the quality and yield of the separation, it is undesirable to have dissolved silica and leached bonded ligand in the final product.
As such, it is important for a packing to be stabile under at least mildly basic conditions because some separations can be greatly improved under these conditions. Under basic conditions, especially over 7.0 pH, silica based packings are not stable because the silica particle simply dissolves away when OH" ions see any exposed silica. The type of ligand attached to the silica greatly affects this stability. Typically, the less polar the ligand on the silica surface, the more stable the packing is to resistant to dissolution. Thus, for example, a CI 8 ligand on the surface provides more protection of the surface than does a C8, CN or C4 ligand.
The poor stability of silica-based bonded phases for LP and HPLC mobile phases having a pH greater than 7 is well-known. Chromatography columns prepared with such packings begin to degrade upon exposure to highly basic mobile phases. Three modes of such failure have been observed: (1) column plugging; (2) loss of peak resolution; and (3) appearance of a large peak at the solvent front of the injection. This sensitivity to high pH exposure is important because many chromatographic materials are often subject to periodic washes with aqueous/organic base solutions. For example, in the commercial chromatographic purification of insulin, the stationary phase of the chromatographic column is washed periodically with dilute aqueous solutions of sodium hydroxide and acetonitrile.
Researchers determined that the type of ligand used and the water content of the mobile phase were found to be important process variables in such bonded phase technologies. In addition, many of these early bonded phases did not receive what is called an endcapping step in which the silica bonded with the major silane is subjected to further reaction with a small chlorosilane such as trimethylchlorosilane. The exposure of the bonded silica to this small silane more completely reacts with the surface and "filled in" sites on the surface which had been left uncovered due to steric hinderances of the major silane not allowing access of the surface. Today, almost all packings are now "endcapped" except those for special applications requiring some "bare" silica showing to achieve good separation.
In addition to normal endcapping, using a single small silane the prior art (Kirkland, et al., J. Chromatogr. A. 762 (1997) 97 - 112) has shown that a double endcapping technique using di and mono functional halosilanes may enhance the stability of bonded chromatographic packings subjected to mobile phases up to pH 10 and 11. This paper supports the thought that this is an important step in
recognizing that details of the bonding may play a significant roll in determining the properties of the bonded phase.
Miller and DiBussolo reported significant differences in a bonded silica resistance to basic pH depending on whether a silica surface was endcapped or not. The endcapped packing having much greater resistance. (Miller, N. and DiBussolo, J. Chromatogr. A. 499 (1990) 317 - 332).
Any enhancement of the lifetime of the packing is a good product improvement. It is therefore the primary focus of the invention to provide a method for increasing the useful life of such silica-based media from an average 750 mL to at least about 1,200 mL of the basic wash solution and preferably 1,500 mL or more (a 50 - 100% improvement in lifetime).
References of Interest
U.S. 4,600,646, Stout
Surface-stabilized silica having methoxy surface coverage sufficient to improve hydrolytic stability.
U.S.5,032,266 Kirkland et al.
Porous silica microspheres having completely silanol-enriched surfaces.
R.K. Her, The Chemistry of Silica, Wiley-Interscience, New York, 1979 pp.40 - 65
Kirkland et al., Stability of silica-based, encapped columns withpH 7 and 11 mobile phases for reversed-phase high performance liquid chromatography, J. Chromatogr. A., 762 (1997) 97 - 112
SUMMARY OF THE INVENTION
In its primary aspect, the invention is directed to a hydroxyl containing porous material having covalently bonded thereon (1) a first silane compound having a chain length in the range of C6 to C20, (2) a second silane compound having a chain length in the range of C2 to C5, and (3) a third silane compound having a chain length in the range of C 1 , wherein said silanes are bonded to the material in a sequence inverse to the number of carbon atoms in the alkyl chain of the silane.
In another aspect, the invention is directed to a method for improving the resistance of a silanized hydroxyl containing material towards basic or acidic solutions, the method comprising covalently bonding to a porous hydroxyl containing material in the following order: (1) a first silane compound having a chain length in the range of C6 to C20; (2) a second silane compound having a chain length in the range of C2 to C5; and (3) a third silane compound having a chain length in the range of C 1.
DETAILED DESCRIPTION OF THE INVENTION
The invention is directed to a hydroxyl containing porous material having covalently bonded thereon (1) a first silane compound having a chain length in the range of C6 to C20, (2) a second silane compound having a chain length in the range of C2 to C5, and (3) a third silane compound having a chain length in the range of C 1 , wherein said silanes are bonded to the material in a sequence inverse to the number of carbon atoms in the alkyl chain of the silane.
It is preferred that the hydroxyl containing porous material be a silica-based, endcapped silica chromatographic media, which are more stable to aqueous organic base solutions having a pH of above 7.
In an alternative embodiment, the silica chromatographic media comprises finely divided, porous silica particles having surface hydroxyl groups, which have been covalently bonded with (1) an alkyl halosilane compound having a chain length of C6 - 20 (2) one or two alkyl halosilane compounds having a chain length of C2 - 5, each of the halosilanes having been individually bonded to the silica to the point of steric saturation in a sequence inverse to the number of carbon atoms in the alkyl chain (i.e. the longest chain bonded first).
It will be recognized that alkyl halosilanes with various halogen entities can be used in the invention. However, chlorides are preferred because they are more economical and more widely available. Suitable catalysts to promote the bonding reactions are basic organic catalysts, particularly organic N-containing bases, which are soluble in the solvent(s) used for the bonding reactions. Imidazole is quite useful for this purpose.
In a preferred embodiment, the chromatographic media of the invention are made by the following process:
A. Finely-divided particles of a porous, rehydroxylated silica are dispersed in inert organic solvent in which the alkyl halosilanes are soluble; B. Along with an effective amount of basic organic catalyst, a first alkyl halosilane having an alkyl ligand of C6 - 20 carbon atoms in length is added to the dispersion and dissolved in the solvent medium in molar excess of the amount required for steric saturation of the hydroxyl groups on the silica surface and heated with reflux to effect covalent bonding to the hydroxyl groups on the silica surface until further bonding of the silane ceases because of steric saturation.
C. A second dimethyl alkyl halosilane having an alkyl ligand of C2 - 6 carbons, which is at least two carbons shorter than the alkyl ligand of the first halosilane, is added to the dispersion from step B and dissolved in the solvent medium in molar excess of the amount required for steric saturation of the hydroxyl groups on the silica surface and heated with reflux to effect covalent bonding to the hydroxyl groups on the silica surface until further bonding of the second silane ceases because of steric saturation.
D. A third alkyl halosilane, having an alkyl ligand, which is at least two carbons shorter than the alkyl ligand of the second halosilane, is reacted with the silanized silica from step C in the same manner prior to separation of the fully silanized silica from the reaction dispersion.
E. The silanized silica from step D is separated from the solvent by filtering the dispersion; and F. Residual unbonded silanes are removed from the separated silica particles by washing them with an aqueous solution of organic solvent in which the silanes are soluble.
As an example of this preferred embodiment, specifically step D, a third silane having a ligand of only 1 carbon atom is added after bonding of the silica with the second silane having a ligand of 4 carbon atoms.
Substrate Hydroxylation:
It is preferred that the surface of the silica substrate used for endcapping by silanization contain about 8.0 micromoles/m of silanol (Si-OH) groups, which is approximately the maximum effective concentration of silanol groups. Above this concentration, a substantial number of silanol groups become buried below the
silica surface. Though such buried silanol groups are detected by thermogravimetric analysis (TGA), they are not available either for chromatographic interactions or for reactions with silanizing agents.
The concentration of silanol groups on the surface of the substrates used in the practice of the invention can be measured by several procedures, including infrared spectroscopy, solid-state magic angle spinning nuclear magnetic resonance, proton spin-counting NMR and thermogravimetric analysis.
If the silica material selected for use in the invention is deficient in hydroxyl groups, i.e. the concentration of surface silanol groups is too low, the material can be hydroxylated by treatment with a dilute aqueous solution of either acid or base. Suitable acids include HF and HNO3 and suitable bases include ammonium hydroxide, sodium hydroxide and organic amines. Hydroxylation or rehydroxylation, as the case may be, can be carried out effectively at temperatures of 25 - 100C for sufficient time to generate the desired concentration of silanol groups.
Steric Saturation:
As used herein, the term "stetric saturation" refers to the maximum concentration of covalent bonding which is obtained with a given silanizing agent before steric hindrance of the alkyl ligand precludes further covalent bonding between the silane and the hydroxyl groups on the silica substrate.
The advantageous properties of this invention can be observed by reference to the following examples which illustrate the invention.
Examples
Example # 1
Preparation of prior art chromatographic media
"Standard" bonded silica having a single endcapping step. Such standard bonded silica is known and presently used in the prior art.
A commercial C-8 packing having a surface bonding of λ.-octyldimethylsilyl followed by a single trimethylsilyl endcapping (BTR LPlOO/10 C8) was used. The packing had a surface area of 342 M2/gram, a nominal pore size of 100 A and a nominal particle size of 10 microns. The analysis showed the packing to be %C =12.6, %H=2.4, %N=0.11. The calculated surface coverage based on C8 is 3.79 uMole of C8/M2 of surface area.
Example # 2
Preparation of "experimental" bonded silica having a double endcapping step
38 grams of vacuum dried BTR LP 100/10 bare silica having a surface of 304 M2/gram, pore size of 110 A and a nominal particle size of 10 microns (BTR Separations, Wilmington, DE) was placed in a multinecked 500ml round bottom flask along with 225 ml of toluene, 29 grams of imidazole, and 56 mL of dimethylformamide. A slow nitrogen purge was used throughout the bonding process. The mixture was stirred under reflux to remove the water by means of the azeotrope along with about 25 mL of toluene. The mixture was cooled below reflux and the moisture trap was replaced with a condensor and 22.0 grams of n- octyldimethylchlorosilane (Silar Labs, Inc., Schenectady, NY ) was added to the flask and the mixture was again brought to reflux temperature. The reflux was continued for 2 hours at which time the mixture was allowed to cool and 16.8 grams of n-butyldimethylchlorosilane(United Chemical Technologies, Bristol, PA) was added to the flask. The flask was brought to reflux and continued to reflux for another 2 hours. The mixture was cooled below reflux and 1 1.8 grams of trimethylchlorosilane(TMCS) (Silar Labs, Inc., Schenectady, NY) was added to the flask. The mixture was again brought to reflux and refluxed for 2 hours. The mixture was allowed to cool below reflux at which time it was filtered hot through a sintered glass funnel. The filter cake was washed with THF/water, filtered to dryness on the filter and then resuspended in THF/water solution in the bonding flask. The mixture was brought to reflux and allowed to cool. At about 40 deg. C. the mixture was filtered, washed with THF and then resuspended in THF and again brought to reflux for 15 minutes. The mixture was then filtered hot, washed with THF and drawn dry on the filter. The free flowing bonded silica was then placed in a vacuum oven @ about 100 deg. C. for 6 hours. Analysis of the dried material yielded the following results: %C=11.37, %H=2.24, %N=.05. This corresponds to a specific coverage value of 3.74 uMole of C8/M2 of surface area.
Packings Tested
1. C8 from Example 1 (prior art) 2. C 8 as set forth above (the invention)
Test Conditions and Protocol
The packings to be tested were loaded as a slurry into standard 0.46 x 25 cm SST columns under high-pressure using standard loading techniques. The columns were initially tested for performance using a sample consisting of uracil, phenol, dimethylaniline, and toluene. The loading conditions were not fully optimized but enough columns were loaded to obtain reasonable plate counts and asymmetry values. Generally this meant 9000-11000 plates and a 10%) peak asymmetry of 0.97-1.2. The columns were then subjected to a basic mobile phase containing 85%o 0.04 M sodium hydroxide in water/ 15%> acetonitrile (176 ml of acetonitrile were added to 1000 ml of water containing 1.6 gm of sodium hyroxide). Typical flow rates of 1.0 -1.5 ml/min used. Most times the columns were flushed with basic mobile phase in paired columns plumbed in parallel on pumping stations consisting of either Perkin Elmer or Thermo Separations isocratic pumps. The eluents from the two columns were collected in separate bottles so the volume of liquid passed through each could be accurately measured.
At certain intervals (typically 200-300 ml), the columns were removed from the pump and tested for performance. These intervals varied based on experience with the different packings.
After the columns were removed from the pumping station they were flushed immediately with water to reduce the pH. Then they were flushed with acetonitrile until a stable baseline was obtained. Then the column was equilibrated with mobile phase for testing.
Testing Performance of Column Using Insulin as a Sample
The test instrument was a Rainin model Dynamax SD-1 HPLC with high pressure mixing gradient elution and a multiwavelenth detector(Rainen model UV-1 detector). Solvents were degassed using an ultrasonic bath prior to use.
The test conditions were:
Detector wavelength=280 nm Flow rate=l .0 ml min. Temperature=ambient (approx. 22 deg. C.) Mobile phase: Solvent A=0.1 M morpholine/0.1 M acetic acid, pH=7.75
Solvent B = acetonitrile. Gradient: 24% B to 35% B in 30 minutes.
Modes of Failure
Failure in the insulin test was defined as:
1 ) Failure to see the impurity peak before or after the main peak,
2) The appearance of a large baseline disturbance or peak following the solvent front.
3) Excessive backpressure of the column indicating collapse of the packing bed.
The second mode was the most common.
Results
Packing Type Volume of hydroxide solution to failure
C8 from Example 1 Lot A 900 mL
C8 from Example 2 Lot B 1500 mL
Lot C 1543 mL