Characterization of chemically bonded phases on precoated thin-layer plates by 29Si and 13C cross-polarization and magic-angle spinning NMR

(1)

Characterization of chemically bonded phases on precoated

thin-layer plates by 29Si and 13C cross-polarization and

magic-angle spinning NMR

Citation for published version (APA):

Haan, de, J. W., Ven, van de, L. J. M., Vries, de, G., & Brinkman, U. A. T. (1986). Characterization of chemically

bonded phases on precoated thin-layer plates by 29Si and 13C cross-polarization and magic-angle spinning

NMR. Chromatographia, 21(12), 687-692. https://doi.org/10.1007/BF02313680

DOI:

10.1007/BF02313680

Document status and date:

Published: 01/01/1986

Document Version:

Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can be

important differences between the submitted version and the official published version of record. People

interested in the research are advised to contact the author for the final version of the publication, or visit the

DOI to the publisher's website.

• The final author version and the galley proof are versions of the publication after peer review.

• The final published version features the final layout of the paper including the volume, issue and page

numbers.

Link to publication

General rights

Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain

• You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:

www.tue.nl/taverne

Take down policy

If you believe that this document breaches copyright please contact us at:

openaccess@tue.nl

providing details and we will investigate your claim.

(2)

Characterization of Chemically Bonded Phases on Precoated Thin-layer

Plates by 29Si and 13C Cross-Polarization and Magig-Angle Spinning NMR

J. W.

de Haan*

/ L. J. M.

van de Ven

Laboratory of Instrumental Analysis, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The Netherlands

G. de Vries / U. A.

Th. Brinkman

Department

of Analytical Chemistry, Free University, De Boeletaan 1083, 1081 HV Amsterdam, The Netherlands

Key Words

Thin layer chromatography Chemically bonded phases

Examination by 29Si and 13C NMR

Summary

The

paper reports results for commercial precoated plates and home-made bonded phases for TLC examined by 29Si and 13C NMR of solid samples.

The NMR spectra readily reveal differences in degree of silylation, details pertaining to the type of reactive silanes used, and differences in the methods of synthesis. They can also be used to explain the cause of the wet- lability problems encountered with some types of (precoated) TLC plates.

of elution of the analytes. It is obvious that a knowledge of the chromatographic process is particularly important in studies on the transfer of data from TLC to column liquid chromatography (CLC), and vice versa, and in work on the prediction of, e.g., octanol/water partition coef- ficients of toxicologically relevant compounds from TLC and/or CLC data [1,2].

More detailed information on the nature of modified silica may be obtained by 29Si and 13C cross-polarization and magic-angle spinning nuclear magnetic resonance (CP- MAS NMR). Several papers have been published for both newly synthesized and commercially available packing materials used in gas chromatography and CLC [ 3 - 7 ] . No work is known, however, on bonded phases coated onto TLC plates. The present communication reports CP-MAS NMR results for eleven types of commercially available precoated plates and three home-made chemically bonded phases for T LC.

Introduction

During recent years, much attention has been devoted to thin-layer chromatography (TLC) on chemically bonded phases, and chromatoplates precoated with non-polar alkyl-modified silicas as well as more polar phenyl-, cyano- and amino-bonded phases are now commercially available from several manufacturers. Although a rapidly increasing number of papers is being published in this area, detailed information on the nature of the precoated bonded phase often is rather inadequate. This is an unfortunate situation, because there are well documented differences between precoated plates offered for sale by different manufac o turers, both as regards the dependence of migration speed on mobile phase composition and, occasionally, the order

Experimental

The

29Si and 13C CP-MAS NMR spectra were run on a Bruker (Karlsruhe, FRG) CXP-300 spectrometer at 59.13 MHz and 75.48 MHz, respectively [5, 7].

200rag of each of the sorbents were scraped from a thin- layer plate and used as such for the CP-MAS NMR studies. In a few preliminary experiments, NMR spectra were run for sorbents removed from a dry - i.e., as received -- precoated plate and sorbents scraped from a plate, that had first been developed with a methanol-water mixture. No noticeable differences were observed, and dry plates were used in all further work. Details on the various types of chromatoplates are given below.

(3)

Results and Discussion

General Characteristics of Precoated and Home-made Plates

Eleven types of commercially available plates were studied, namely nine non-polar plates precoated w i t h C1-C18- modified silicas, and two polar plates, an amino- and a cyano-bonded phase. Data on plate designation, manufacturer, carbon chain length and approximate mean particle diameter are summarized in Table I. Remarks about the silanization procedure are partly taken from the published literature [1, 2], and partly based on information received from the manufacturers. The following remarks should be made.

(i) Most plates are of a normal T L C quality with, generally, a 1 0 - 1 5 # m particle size range. Exceptions are the HPTLC-quality RP-coated and cyano- and amino- bonded Merck plates, which feature the 5 - 7 # m particle size diameter typical for high-performance TLC.

(ii) Apart from the difference in particle size, the main difference between the HPTLC- and TLC-quality RP- coated plates from Merck is that, w i t h the TLC-quality layers, silanization is what can best be described as marginally less than exhaustive. No quantitative data are, however, known.

(iii) With the SIL C18 plates, the manufacturers state that they are either 'completely' or '50%' silanized. Un- fortunately, they do not state what this percentage refers to, but it is unlikely that it represents the percentage of (accessible) silanol groups reacted.

The home-made plates were prepared by Ericsson, using in situ modification of thin-layer plates precoated with silica [8]. Modification was done with either trichlorooctadecylsilane or monochlorodimethyloctadecylsilane in dry toluene, with reaction times of 1 2 - 2 4 h , and reaction temperatures of 25 or 120~

298i and 13C CP-MAS NMR spectra

The information to be obtained from 2esi and 13C NMR spectra of modified silica materials for chromatography can be qualitative and quantitative in nature [4].

29Si NMR yields three distinguishable signals for silica at - 9 2 , - 101, and - 1 1 0 p p m with respect to tetramethyl- silane (TMS). These can be assigned unequivocally to 29Si nuclei in silanediol, silanol and 'quaternary' groups, respectively (see Fig. 1 and the - 1 0 1 and - 1 1 1 p p m peaks in the spectra shown in Fig. 2 below).

After silylation with a monofunctional silane, an ad- ditional signal is found near +12ppm, which belongs to the silane 29Si nuclei (Fig. 2A; only monodentate linkages are formed). Upon silylation w i t h bi- or trifunctional silanes, mono-, bi- and, in the latter case, tridentate linkages and cross-polymerized materials can be formed. The positions of the corresponding 29Si NMR signals depend to a certain extent on the presence or absence of residual reactive moieties on the silane and, in the former case, also on their nature. A n example of a silica derivatized with a bifunctional silane is shown in Fig, 2B. Derivatization w i t h a trifunctional silane may yield a spectrum like that shown in Fig. 2C. In both cases, the peak assignments are given in the legend to the figures. Not all differences in situations can be distinguished. This is particularly true for

29Si

nuclei involved in tridentate linkages; here, at present, no distinction can be made between silane to surface and silane to silane (cross-polymerization) linkages. On the other hand, mono- and bidentate linkages formed from bifunctional silanes can be distinguished. The same applies for mono-, bi- and tridentate linkages formed from trifunctional silanes,

-110

-101

- 9 2 ppm

OH

HO OH

I

\/

Si

/ / \ \

/1\

/\

Q O ~ / ~ O O

Q Q

Fig. 1

A silica surface with a silanediol, a silanol and a 'quaternary' group which show 29Si signals at -92, -101 and- 110ppm, respectively.

Table I. Commercial chemically bonded TLC plates

Manufacturer Plate designation Carbon chain dp ( # m } Silanization Opti-UP Antec, Bennwil, Switzerland Macherey-Nagel, D0ren, FRG Merck, Darmstadt. FRG Whatman, Clifton, N J, USA SIL C18-100 and C18-50 Silanized silica T LC-quality RP-8 & -18 HPTLC-quality RP-2, -8, -18 HPTLC NH 2 HPTLC CN KC18 12 18 2 X 1 8,18 2 X 1 8 & 1 8 18 10--20 5--10 11--13 11-13 5--7 5--7 5--7 10--14 or 20 C12-trichlorosilane 100 and 50% silanization di-C1 -dichlorosilane silanization not completely exhaustive

"y-aminopropyl group ~'-cyanopropyl group end-capped with C2

(4)

\

b)

I I I I | I L I I ! I I I I I

c -50 -lOO 6

Fig. 2 A

Typical 29Si CP MAS NMR spectrum of a silica, derivatized with a monofunctional silane (monochlorodimethyloetadecylsilane; M. Ericsson). The assignments are:

-- + 13ppm: (SiO)29Si(CH3)2C18H37; -- - 101 ppm: (SiO)329SiOH;

-- - 111 pprn: (Si0)429Si.

Fig. 2B

Typical 29Si CP MAS NMR spectrum of a silica C18, derivatized with a bifunctional silane (TLC-quality RP-18; Merck). The assignments are:

-- - 7.5ppm: (SiO)29Si(CH3) (OCH 3)R; - - - 16 ppm:(SiO)229Si(CH3)R; - - 1 0 1 ppmand-111ppm:see2A.

Fig. 2 0

Typical 29Si CP MAS NMR spectrum of a silica 018 , derivatized with a trifunctional silane (trichlorooctadecylsilane, 25~ M. Ericsson). The assignments are:

57 ppm: (SiO)229Si(OH) R; -- - 67ppm: (SiO)329SiR;

-- - 101ppm a n d - 111 ppm; see 2A.

which resonate in a spectral region that is well separated from that of the products of bifunctional silanes. This also implies that the use of mixed reagents, e.g., a bi- and a t r i f u n c t i o n a l silane, can be deduced from the spectra. Finally, end-capping can be found by comparing results of 29Si and 13C NMR.

130 NMR can be used to check the presence of residual carbon-containing reactive groups after silylation (e.g., a l k o x y groups). Further, functional groups on the silane chain can either by identified directly (e.g., cyano groups) or indirectly by their influence on the 130 NMR chemical shifts of carbon atoms (e,g., amino groups). For chains w h i c h contain no substituents beyond the silane silicon atom, a m a x i m u m of about 8 different signals can be observed under favourable conditions. For longer chains such as C18 chains, signal overlap occurs.

A l l spectra in th'is study have been obtained by means of the CP excitation method. The use of pulse excitation f o r 29Si MAS NMR spectra was advocated recently [9], but the measuring times required to obtain useful signal-to- noise ratios w o u l d have been u n d u l y long for our samples. On the other hand it is unfortunate that in contrast w i t h pulse excitation, CP excitation depends on several variables. Because of this, quantitative studies are d i f f i c u l t , and cali- bration experiments are usually necessary.

In the course of this study we performed a number of preliminary measurements of CP curves [10]. We observed that the differences in CP characteristics f o r different silicas are usually negligible. However, different degrees of silanization w i l l certainly cause a loss of accuracy when o n l y one set of measurement conditions is used, as in the present case. We estimate that the m a x i m u m deviations w i l l amount to 5 - 7 % (relative). The reproducibility of the measurements is o f the same order; so that a relative inaccuracy of ca. 15% is the best that can be claimed. The 29Si CP-MAS spectra recorded for 14 different chem-

ically bonded phases have been used to calculate degrees of silanization and product distributions. The results are summarized in Table II; one should note t h a t the defini- t i o n of coverage differs from that used in many other studies (/lmole/area unit). Relevant aspects are further discussed in the next sections.

C o m m e r c i a l Non-polar Precoated T L C P l a t e s

M e r c k : The results for the f o u r types of Merck RP plates indicate that for the RP-8 as well as the RP-18 phase silanization has progressed about 50% further for the HPTLC as compared to the T L C plates. This results in a higher coverage and, thus, creates the w e t t a b i l i t y problems w i t h eluents containing over 3 0 - 4 0 % of water well k n o w n for HPTLC-quality, but not for T L C - q u a l i t y Merck plates [1, 2]. The difference in coverage between the HPTLC and T L C phases appears to stem largely from differences in concentration of the bidentate linkages, while the concentrations of the monodentates are the same w i t h i n the experimental limits.

It is interesting that the differences in coverage between the RP-8 and RP-18 materials, for the TLC- a n d the HPTLC- quality phases, are considerably smaller than is usually f o u n d between octyl- and octadecyl-bonded phases for CLC from the same manufacturer. For the CLC phases, the coverages are generally larger than for the T L C phases, especially in the case of the octyl phase. The 29Si CP-MAS NMR spectrum of LiChrosorb RP-8 for CLC shows o n l y

(5)

T a b l e II. Percentage distribution of total 29Si MAS NMR signals* NMR signals Sample Commercial RP-18 HPTLC RP-18 TLC RP-8 HPTLC RP-8 TLC Sil. silica KC18 Opti-UP C 1 2 S I L C18-50 SIL C18-100 CN HPTLC NH 2 HPTLC Home made Tri-CI-C18 (120~ Tri-CI-C18 (25~ Mono-CI-C18 Monodent. bident. 7 28 7 15 1 0 27 8 17 - - 50 - - 8 - - 1 0 - - 2 5 - - 2 7 - - 2 7 - - 15 11 18 - - 27 22 trident. w B 8 5 5 18 35 7 15 silanediol 4 4

* Excluding those of the quaternary silicon atoms in the silica lattice. ** Number of silanized silanol groups/initial number of silanol groups (in %).

Per cent coverage** silanol 65 35 78 22 59 37 71 25 50 50 84 16 85 1 5 70 30 55 45 73 27 50 50 64 36 58 42 78 22

a very small signal of residual silanol groups, whereas distinct signals are observed for all T L C phases investigated in the present study. Residual silanediol signals were also found for both the TLC- and the HPTLC-quality RP-8 phase.

In general, differences in coverage between octyl- and octadecylsilane-type phases - all other factors remaining equal - can be understood in terms of the larger effective area of longer chains, which results in larger steric inter- actions and, hence, in lower coverage for a given case. The differences noticed here between phases, used for CLC and (HP)TLC must, however, originate from differences in the method of synthesis. Whereas the T L C phases are prepared starting with SI-60 silica and dichloro- alkylmethylsilanes, the CLC phases are produced by silylation of S1-100 silica with dialkoxyalkylmethylsilanes. Furthermore, the reaction temperatures are different. The differences between HPTLC- and TLC-quality phases can conceivably be due to either a somewhat higher reaction temperature or to the presence of some residual water during silylation for HPTLC phases.

Finally, the silanized silica plate is another Merck product marketed around 1970 and, in a way, the TLC-quality pendant of the HPTLC-quality RP-2 plate (which was not included in the present study). The degree of silylation of the silanized silica is unusually high: the coverage is about 50% and, virtually all silane is present as a bidentate linkage; 13C NMR analysis revealed only one absorption at - 1 . 5 p p m . This points to the presence of dimethyl- disiloxysilane groups at the surface of these materials and confirms the statement in Table I, that the silanized silica (and the LiChrosorb RP-2 phase) is a di-C1 material rather than a C2 material as suggested by the '2' designation.

Whatman and Antec: Comparison of the 29Si NMR spectrum of the Whatman KCt8 sample w i t h those of the four Merck RP samples, shows that the Whatman sample has

been prepared using a trifunctional silane, which yields about equal amounts of bi- and tridentate linkages. The overall coverage is significantly lower than with any of the RP-8 and RP-18 samples. A similar statement holds for the Antec Opti-UP C12 sample, but in this case the silanol signal is so strong (compared with, e.g., the quaternary 29Si signal) that the original silica must have contained a higher concentration of surface silanol groups or, less probably, that the reaction conditions have been unusually mild.

The relatively low coverages of the K018 and C12 samples at least partly explain w h y migration speed on the What- man and Antec plates often is distinctly higher than on the other non-polar chemically bonded phases [1,11].

Macherey-Nagel: Study of the two samples from Macherey- Nagel, SIL C18-50 and SIL C18-100, showed that both phases were prepared using trifunctional silanes: only bi- and tridentate linkages are present with 29Si NMR signals a t - 5 9 p p m and - 6 6 p p m , respectively. A much larger concentration of tridentates is mainly responsible for the larger overall degree of silylation of the SI L C18-100 sample. This causes serious wettability problems which are largely absent in the case of the less exhaustively silanized SIL C18-50 phase, as demonstrated in, e.g., refs. [1] and [11].

C o m m e r c i a l P o l a r P r e c o a t e d T L C P l a t e s

The cyano- and amino-bonded phases, both produced by Merck, allow one to make an interesting comparison. The cyano phase shows a coverage of about 27% which is all bidentate, with no monodentate signal showing up. The amino phase, however, apart from having a much higher coverage, displays bidentate (15%), and tridentate and/or cross-polymerization (35%) signals. The differences between these, in principle, rather closely related polar bonded phases can be traced to their methods of preparation.

(6)

Whereas the cyano phase is manufactured with methyl- dichloro-3'-cyanopropylsilane, the amino phase is synthesized by means of triethoxy-D'-aminopropylsilane. The NMR spectrum of the amino phase shows characteristics which satisfactorily agree with results obtained by Koenig and coworkers [12] and by our group [13]. The shielding of the bidentate 29Si NMR signal of the amino phase with respect to the other bidentates reported in the present study, originates from the presence of a methyl group (instead of chloro, alkoxy or hydroxy groups) on the Si nucleus in question.

The differences in per cent coverage of the cyano and amino phase (as defined in this study; cf. Table II) can not be related directly to the manufacturer's data on coverage in /~mol/m 2 which show a higher coverage for the cyano phase.

H o m e - m a d e B o n d e d Phases

Two precoated plates were sent to us by M. Ericsson (Stockholm, Sweden, cf. ref. [8]) which had been prepared by in situ coating of precoated silica thin-layer plates with trichlorooctadecylsilane as reagent; the reaction temperatures were 120~ and 25~ respectively. Intuitively, one might expect a higher coverage and higher percentage of bi- and tridentates for the 120~ sample. This is, however, not the case (see Table II). In the 120~ sample even monodentate linkages are still observed, while the overall coverage is clearly lower than for the 25~ sample. An ex- planation for this apparent anomaly can be found in the fact that the 25~ sample was prepared in an atmosphere of ca. 15% relative humidity. This causes relatively easy hydrolysis of the starting silanes, and giyes rise to more bl- and tridentate formation and to a higher coverage. It may also explain why, even with a higher overall coverage this sample has the higher silanol to quaternary silicon ratio (hydrolysis of siloxane bridges for the 25~ sample during silylation; data not shown in this paper).

In contrast with the plates described above, the third sample provided by M. Ericsson was prepared using a monochlorosilane reagent (24h reaction at 120~ This can be deduced from the presence of only one 29Si signal near + 13ppm. In spite of the rather long reaction time, the degree of silylation for this sample is lower than that of the samples produced by reaction with the trichlorosilane reagent.

Some chromatographic experiments were carried out with the three types of home-made plates and, for comparison, with TLC-quality RP-18 plates. Six phthalate esters, ranging from C2 to C13, were used as analytes, and methanol-water (90:10) and methanol-water (75:25) were the eluents; the results are shown in Fig. 3. It is evident that the water- richer eluent provides more unambiguous information about the potential of the various thin layers. Wettability problems occur rather rapidly with both phases prepared using the trichlorosilane reagent, and are most serious with the phase displaying the highest percentage of cQverage. The monochlorosilane-based phase is much less hydrophobic - w h ] c h again agrees with expectations based on the data in Table II - and should obviously be preferred.

E r i c s s o n - C1B-RPTLC- phthalates methanol 9 0 % 131 281 281 19 ~ 0 0 0 o 0 ₀ Q o 0 o 0 o o (3 O O o o o o o o o o R P - 1 8 1 2 3 methanol 75~ 17' 60' l h O O lh O C) o 0 o 0 R P - 1 8 1 2 3 Fig. 3

Thin-layer chromatography (5-cm run) of a mixture containing six phthalate esters using (left) methanol-water (90: 10) and (right) methanol-water (75:25) as eluent. Times of run are indicated in the figure. Thin layers: (RP-18) MerckTLC-quality RP-18;Ericsson: (1) trichlorosilane, 12h at 120~ (2) trichlorosilane, 12h at 25~ 14% rel. humidity pretreatment, (3) monochlorosilane, 24h at 120~

Even so, it is much more hydrophobic than the TLC-quality RP-18 precoated plate (60 vs. 17min migration time with the 75:25 methanol-water mixturet) and, actually, displays a behaviour more characteristic of a HPTLC-quality RP phase. As regards chromatographic resolution, the phthalate ester separation is fully comparable to that achieved on the commercial plate.

C o n c l u s i o n

The present study demonstrates that high-resolution NMR, in particular 29Si CP-MAS NMR, can give valuable information on the nature of precoated stationary phases in thin- layer chromatography. The characteristics of new types and brands of (precoated) thin-layer plates can easily be compared with those of (HP)TLC plates already commercially available, and the results so obtained can confirm and extend information gained from chromatographic studies. Comparison of NMR data recorded for stationary phases utilized in thin-layer and in column liquid chromatography is another fruitful area of application. In summary, high-resolution NMR is rapidly becoming an in- dispensable technique for all workers who have to deal with the preparation, characterization and intercom- parison of chromatographic packing materials, and with the interpretation of experimental results obtained with such materials.

A c k n o w l e d g e m e n t s

We thank Dr. M. Ericsson (Stockholm, Sweden) for providing us with several divergent types of home-made precoated plates, and Dr. W. Jost (Merck, Darmstadt, FRG) for sending us useful information.

(7)

References

[1] U.A. Th. Brinkrnan, G. de Vries, J. Chromatogr., 265, 105 (1983).

[2] U,A. Th. Brinkman, in: Instrumental HPTLC, WL~rzburg 1985 (ed., R. E. Kaiser), Institute for Chromatography, Bad OLirkheim, FRG, 1985, p. 25.

[ 3 ] G. E. Maciel, D. W. Sindorf, V. Bartuska, J. Chromatogr., 205, 438 (1981).

[4] D.W. Sindorf, G.E. Maciel, J. Am. Chem. Soc., 105, 1848 (1983) and references cited therein.

[5a] G. Rutten, A. van de Ven, J. de Haan, L. van de Ven, J. Rijks, J. High Res. Chromatogr. & Chromatogr. Commun., 7, 607 (1984).

[5b] G. Rutten, J. de Haan, L. van de Ven, A. van de Ven, H. van Cruchten, J. Rijks, J. High Res. Chromatogr. & Chromatogr. Commun., 8,664 (1985).

[6] E. Bayer, K. Albert, J. Reiners, M. Nieder, D. M6ller, J.

Chromatogr., 264, 197 (1983).

[7a] H.A. Claessens, L . J . M . van de Ven, J, W. de Haan, C.A. Cramers, N. Vonk, J. High Res. Chromatogr. & Chromatogr. Commun., 6,433 (1983).

[7b] H.A. Claessens, C.A, Cramers, J. W. deHaan, F . A . H . den Otter, L. J. M. van de Ven, P. J. Andree, G. J. de Jong, IV. Larnmers, J. Wijma, J. Zeeman, Chromatographia, 20, 582 (1985).

[Sa] M. Ericsson, L. Blomberg, J. High Res. Chromatogr. & Chro- matogr. Commun., 3, 345 (1980).

[8b] M. Ericsson, L. Blomberg, J. High Res. Chromatogr. & Chro- matogr. Commun., 6, 95 (1983).

[9] C.A. Fyfe, G. C. GobbL G.J. Kennedy, J. Phys. Chem., 89, 277 (1985).

[10] O.W. Sindorf, G.E. Maciel, J. Am. Chem. Soc., 102, 7606 (1980).

[11] U. A. Th. Brinkrnan, G. de Fries, J. Chromatogr., 258, 43 (1983).

[12] C.H. Chian9, N.I. Lin, J.L. Koenig, J. Colloid Interface Sci., 86, 26 (1982).

[13] J.W. deHaan, H.M. vandeBogaert, J.J. Ponj#e, L . J . M . van de Ven, J. Colloid Interface Sci., 110, 591 (1986).

Received: July 14, 1986 Accepted: Sept. 23, 1986 E