The Four Pillars of Welch Materials ODS Columns

When industry veterans mention the “four pillars”, many will immediately think of the four gold-hued chiral columns from Daicel — and rightly so; those products opened a new era for chiral separations.

On the contrary, conventional octadecylsilane (ODS/C18) columns may seem commonplace — nearly every manufacturer offers several variants. Who in this field has not used one ODS column?

However, there is still a question that is always worth a thought: what makes an ODS column genuinely “good”? From the early days of Welch Materials being the only manufacturer in China to the current abundance of options, we have kept ourselves pondering this question, and to answer it, we have asked a huge quantity of users, summarizing their wish list as follows:

• Tolerate higher column temperatures and more extreme pH conditions.
• Provide appropriate retention for a wide range of compounds.
• Deliver longer service life.
• Produce good peak shapes regardless of sample matrix.
• Ideally, separate everything.
• ……

Although we wish to satisfy all the above, we must obey scientific reality, rather than being guided by plainly our superficial familiarity and even misconceptions of ODS columns.

In this article, we introduce four Welch ODS columns — Ultisil XB-C18, Ultisil LP-C18, Xtimate C18, and Ultisil AQ-C18 — and use them as examples to discuss this most-used “companion” and “familiar stranger”. But before continuing, please note that the following draws on our long experience and may conflict with some conventional views; critique and correction are welcome.

It is a common practice that, as molecules are becoming more and more complex, both manufacturers and users are pursuing “innovative” stationary phases to deal with their analyses. While this is positive for the industry, it has also led astray.

Despite the limitations, “old-fashioned” ODS columns remain the most commonly used tool due to they being sufficiently simple and pure. Their dominant interaction mechanism is straightforward, which enables many broadly applicable, effective methods for routine use. Experienced chromatographers can often predict reasonable mobile-phase choices and retention behavior from a compound’s structure — something that becomes harder with more polar stationary phases.

And this is why these four purest ODS designs — without fancy embedded polar groups or exotic endcapping — are taken into discussion deliberately.

• Ultisil XB-C18 — the most plain by design.
• Ultisil LP-C18 — intended for use with strongly acidic mobile phases.
• Xtimate C18 — the preferred choice for basic mobile phases.
• Ultisil AQ-C18 — the first choice when using pure water mobile phases.

These descriptors reflect both our marketing and how most users apply them. But does that exhaust their meaning?

From the characteristics above, one might infer that the most “ordinary” ODS cannot tolerate strong acid, base, or pure water. And let’s analyze each issue in turn.

Bonding an octadecylsilane chain to silica replaces a surface silanol proton with an organic chain. This reaction is chemically reversible, and during manufacture we drive the reaction conditions toward a desired dynamic equilibrium to produce high-performance and stable packings.

In routine use, the bonded phase will gradually hydrolyze; higher acidity and elevated temperature accelerate this loss.

To address this issue, Ultisil LP-C18 was designed. Its main features are:

• The methyl groups on the conventional octadecylsilane side chains are replaced with isobutyl groups, increasing steric hindrance around the primary bonded moiety and improving acid resistance.
• An un-endcapped design eliminates the more labile endcap moieties that can hydrolyze more readily than the primary bonded chains.
• The silica surface receives a re-hydroxylation treatment to improve surface uniformity and thus minimize the performance variability that can be amplified by the un-endcapped design.

Given these design choices, it would be a mistake to limit LP-C18 solely to strongly acidic mobile phases. The unique bonded group results in lower polarity and greater steric bulk; therefore its influence on retention differs from merely changing bonding density.

Moreover, controlled exposure of silica surface silanols should not be viewed solely as a drawback — the silanols can provide useful hydrogen-bonding and ion-exchange interactions under different mobile-phase conditions (functionally similar in some respects to deliberately introduced charged endcapping groups).

Caution: because LP-C18 is un-endcapped, the intrinsic characteristics of the silica substrate are more pronounced than in endcapped products, and its tolerance to salts and bases can be weaker than that of endcapped ODS columns. Careful attention is required during use.

Silica’s vulnerability to base stems from nucleophilic attack by hydroxide ions at the Si–O–Si network, causing hydrolytic cleavage of the silica framework. Thus the fundamental limitation of traditional silica-based ODS in basic mobile phases lies with the silica support itself.

Although surface endcapping can mitigate some surface reactivity, and such products exist on the market, it remains unclear what endcapping level can be truly achievable, even to an experienced column manufacturer like us — we have no reliable way to measure an exact endcap coverage and suspect it remains far from 100%.

An alternative approach is to change the base material. Polymer-based supports are inherently resistant to base, but they bring tradeoffs: lower mechanical strength, reduced column efficiency relative to silica, and solvent-dependent swelling.

This context gave rise to hybrid particles — materials that combine silica and organic polymeric character. Waters’ second-generation hybrid (XBridge) made using basic mobile phases on an ODS column practical in a meaningful way (we heartfeltly salute that contribution!).

However, full hybrid particles were not available to us, and to overcome this problem, we developed a surface-hybridization strategy: graft an organic layer onto conventional silica via organic–inorganic grafting to approximate the properties of hybrid particles. The Xtimate series was created using this approach.

Many will assume a surface-hybrid is an inferior “knock-off” compared with fully hybrid particles. In our view, however, this compromise has merits:

• As this treatment treats surface only, theoretically it can be applied to silica of various pore sizes and surface areas without large process changes;
• Hybrid particles themselves are a silica–polymer composite and their mechanical strength can actually be lower than pure silica, whereas surface-hybridized packings retain the mechanical robustness of traditional silica.
• As to empirical performance, we will not prejudge — real samples provide the final verdict. 

Note: in acidic or neutral mobile phases containing salts, increased ionic strength can accelerate dissolution of conventional silica, shortening column life; in such cases the advantages of hybrid or surface-hybrid particles become especially apparent and are worth trying.

This topic inevitably involves the concept of “phase collapse”. Practitioners vary in their interpretation, and for clarity, our definition of phase collapse is: when a hydrophobic bonded phase is exposed to high proportions of aqueous mobile phase, incompatibility between the two phases can cause the stationary phase to “collapse” (retract) to minimize contact area — much like water beading on a lotus leaf or on waterproof fabric.

This phenomenon is normal, yet it is sometimes treated falsely as catastrophic. Would water-resistant ODS columns face no phase collapse? NO.

Most packings in routine use are porous silica intended to allow full contact among sample, mobile phase, and stationary phase for optimal separations; it means that efficient mass transfer requires the mobile phase to penetrate the pore network. However, incompatibility between phases causes the aqueous mobile phase to be unable to enter the pores effectively, resulting in retention behavior changing.

How to solve it? The solution is not a single trick to a small part of the pores, but an overall change to the pore environment. A wide range of strategies can increase apparent hydrophilicity of the pore environment, including:

• Lowering bonding density;
• Introducing polar groups into the bonded chains or via endcapping (including controlled absence of endcapping);
• Increasing pore size;
• ……

What degree of overall pore polarity yields genuine “hydrophilicity”? we cannot and do not claim to know a precise threshold, nor do we consider it necessary. Nominal pore diameter is an average; individual pore geometry varies, and operational conditions such as temperature and column pressure also influence behavior.

Therefore, it leaves to the empirical observation rather than theoretical pronouncement to answer questions like “what is the highest percentage of water this column can tolerate?”

One aside: for highly polar analytes, a hydrophilic ODS column will not always provide greater retention than a conventional ODS column. When dealing with strongly polar analytes, test multiple columns (hydrophilic and non-hydrophilic; hydrophilic with different designs) to find the right selectivity.