HPLC verification, in plain English

If you have ever looked at a certificate of analysis for almost any modern wellness compound, analytical sample, or reference compound, you have probably seen three letters stamped near the top: HPLC. Sometimes they appear beside a number with a per cent sign. Sometimes beside a wavy line that looks like a polygraph readout. Almost always, the document expects you to nod knowingly and move on.

The trouble is that “HPLC” is a black box for most readers. It is treated as a kind of magic certificate, as if the letters themselves were the guarantee. They are not. HPLC is a method, and like every method it has a particular job it does well, a particular job it does badly, and a long list of things it cannot do at all. Knowing which is which is the difference between reading a chromatogram and being sold one.

This post is an attempt to explain, in plain English, what high performance liquid chromatography actually is, where it came from, what an HPLC trace can and cannot tell you, and why it is the single most useful document you can ask a vendor to produce.

The core idea: separating molecules by speed

Imagine you and a friend agree to walk through a crowded train station to the exit on the far side. You walk straight through. Your friend, who is more sociable, stops to chat with every person they recognise. You arrive at the exit first. Your friend arrives much later. If a stranger at the exit timed both of you, they could tell you apart simply by how long each of you took.

That is, almost exactly, how chromatography works.

An HPLC system has two phases. The mobile phase is a liquid solvent that flows continuously through the instrument, like the crowd moving through the station. The stationary phase is a fine, packed material inside a narrow tube called the column; it is the friend who stops to chat. When you inject a tiny volume of sample into the mobile phase, every molecule in that sample is carried into the column. Once inside, each molecule spends some of its time dissolved in the moving liquid and some of its time “stuck” to the stationary phase. Molecules that interact strongly with the stationary phase loiter; molecules that don’t, race through.

By the time the mobile phase exits the far end of the column, what went in as a mixture has come out as a tidy procession. A detector at the end of the column watches the stream and records what passes by, when. The output, a plot of detector signal against time, is called a chromatogram.

The “high performance” part is mostly about engineering. The column is packed with extremely small, uniform particles, and the mobile phase is pushed through it under very high pressure. Smaller particles mean more surface area for those interactions to happen on, which means sharper separations. The Royal Society of Chemistry’s analytical chemistry teaching materials are a good place to see the basic idea drawn out with diagrams (RSC Education: Chromatography).

A brief history: from leaf pigments to pharma

Chromatography was invented by a botanist who was trying to understand the colours inside leaves. In 1900, a Russian-Italian scientist named Mikhail Semyonovich Tsvet poured an extract of green plant pigments through a glass tube packed with powdered calcium carbonate. As the solvent dripped down, the muddy green smear separated, almost theatrically, into distinct horizontal bands of yellow, orange and green: the carotenoids and the chlorophylls, sorted by how strongly each one clung to the chalk. He called the technique chromatography, from the Greek for “colour writing.” Tsvet first described it publicly at the end of 1901 and in print in 1906 (Mikhail Tsvet, Wikipedia; Britannica: Mikhail Semyonovich Tsvet).

His method was almost lost. A famous chemist of the day tried to reproduce it, used the wrong adsorbent, destroyed his sample, and concluded the technique didn’t work. Tsvet’s chromatography fell into obscurity for two decades before being revived in the 1930s.

The leap from a glass tube of chalk to the steel-and-pump instruments in modern labs was largely the work of a Hungarian-American chemical engineer named Csaba Horváth. Starting in the mid-1960s at Yale, Horváth set out to apply real engineering, high pressures, small particles, careful detectors, to liquid chromatography. The phrase “high pressure liquid chromatography” appears in his work by the late 1960s, and “high performance liquid chromatography,” the term we use today, was popularised at a 1970 conference (Csaba Horváth, Wikipedia; LCGC: Csaba Horváth and the first modern HPLC). Within a couple of decades it had become arguably the most widely used analytical technique in chemistry and biochemistry.

How to read a chromatogram

A chromatogram is a deceptively simple-looking thing. The horizontal axis is time. The vertical axis is the detector’s signal: how much “stuff” is going past at that moment. The baseline is flat and quiet. Each peak that rises above the baseline corresponds to a band of one particular molecule exiting the column.

There are three numbers worth knowing.

Retention time

This is the time, measured from injection, at which a peak appears. Under fixed conditions (same column, same mobile phase, same temperature, same flow rate), a given molecule will always come out at roughly the same time. Retention time is therefore a fingerprint. If you run a known reference compound and see its peak at 4.7 minutes, and your unknown sample also produces a peak at 4.7 minutes under identical conditions, that is strong (though not conclusive) evidence the two are the same substance.

Peak area

The area under each peak is, in well-behaved methods, directly proportional to how much of that molecule was injected. This is what lets HPLC quantify purity. If your sample produces one large peak that accounts for 99.2 per cent of the total area, and a couple of tiny peaks accounting for the remaining 0.8 per cent, you have an estimate of how much of the sample is your target compound and how much is “everything else the detector could see.”

Peak shape

A clean, symmetrical peak suggests a clean separation. A peak that is broad, lopsided, or hides a “shoulder” suggests that two things are co-eluting, coming out together. This is one of the easier things to fudge in a printed certificate, and one of the easier things to spot if you have the original trace.

The international pharmacopoeial standard that defines all this rigorously, including which peaks are allowed to count and how a method has to be validated before its numbers mean anything, is the United States Pharmacopeia’s General Chapter <621> on chromatography, revised most recently in 2022 (USP <621> Chromatography).

What HPLC verifies, and what it absolutely doesn’t

This is the part that gets glossed over most often.

HPLC, properly run with a matched reference standard, can tell you two things with confidence:

• Identity, in the limited sense that a peak in your sample shares a retention time, and ideally a UV absorption profile, with a known reference compound run under the same conditions.
• Purity by area per cent: the fraction of the total detector signal attributable to your target peak versus everything else the detector saw.

Both of those qualifications matter. “Under the same conditions” rules out comparing a vendor’s chromatogram to someone else’s chromatogram run on a different machine with different solvents. “Everything else the detector saw” rules out anything the detector cannot see, which, with a standard UV detector, can be a great deal. Salts, certain solvents, and molecules without the right chemical features are invisible to UV at common wavelengths. A 99 per cent area-per-cent purity result, in other words, is 99 per cent of what the instrument was able to detect, not 99 per cent of what is in the bottle.

And then there is the long list of things HPLC cannot tell you, full stop:

• It cannot tell you whether a substance is safe.
• It cannot tell you whether the substance is biologically active, or active at what level.
• It cannot tell you whether a sample is stable, or how long it will remain so.
• It cannot tell you about contaminants the detector is blind to, including many heavy metals and microbiological contaminants.
• It cannot definitively confirm identity on its own. Two very different molecules can elute at exactly the same time.

This last point is why serious analytical chemists almost never rely on HPLC alone for identification.

UHPLC and what has changed in twenty years

If HPLC is the 1970s sedan of analytical chemistry, UHPLC, ultra-high-performance liquid chromatography, is the modern carbon-fibre version. The chemistry is the same. The engineering tolerances are tighter.

The single biggest change is particle size. Conventional HPLC columns are packed with particles around 3 to 5 micrometres across. UHPLC columns use particles smaller than 2 micrometres. Smaller particles give sharper separations, but they also resist flow much more strongly, which is why UHPLC instruments routinely operate at pressures north of 1000 bar, roughly the pressure at the bottom of a 10-kilometre-deep ocean. The practical payoff is dramatic: runs that used to take half an hour can finish in five minutes, with better resolution between closely related compounds (ScienceDirect: Ultra-High-Performance Liquid Chromatography overview).

For the everyday reader of a certificate of analysis, UHPLC is good news. It means modern labs can resolve closely related impurities that older instruments would have lumped into one fat peak.

When HPLC marries mass spectrometry: LC-MS

The honest limitation of HPLC, that two unrelated molecules can share a retention time, is solved by coupling it to a second instrument that asks an entirely different question: how much does each molecule weigh?

That second instrument is a mass spectrometer. In an LC-MS system, the chromatograph’s job is unchanged: separate the mixture in time. But instead of feeding the separated stream into a simple UV detector, it is fed into an ion source that gives each molecule an electric charge and shoots it into a vacuum chamber, where its mass-to-charge ratio is measured with extraordinary precision. Each peak in the chromatogram is now accompanied by a mass spectrum, a sort of molecular weight signature.

The two methods complement each other almost perfectly. Chromatography says when something arrived; mass spectrometry says what it weighs, and often, with tandem MS, what it is built out of. Together they can confirm identity in a way neither can manage alone, which is why LC-MS has become the default tool for serious analytical work in pharmaceuticals, clinical labs, and research (Pitt, J.J. (2009). Principles and Applications of LC-MS in Clinical Biochemistry. Clinical Biochemist Reviews 30(1); RSC: Guide to Reliable Quantitative LC-MS Measurements).

If a vendor’s documentation includes an LC-MS trace alongside a plain HPLC trace, that is a meaningfully stronger document than HPLC alone.

Why an HPLC trace is the single most useful document you can ask for

HPLC is not a guarantee of safety. It is not a guarantee of biological effect. It is not even a guarantee of identity in the strictest sense. What it is is the cheapest, most universal, most widely understood snapshot of “what is in this bottle, in roughly what proportions, as far as a UV detector is able to tell.” For research-grade material and wellness compound testing, that is genuinely useful information, especially when you have the original chromatogram rather than a one-line summary.

If you only have time to ask a vendor for one document, ask for the full HPLC trace, not just the certificate of analysis. Look at the baseline (is it clean and flat, or noisy?). Look at the main peak (is it tall, sharp, symmetrical?). Look at the smaller peaks (are there many of them, or just a few faint ones?). Look at the conditions (column type, mobile phase, detection wavelength) printed on the report. Compare the retention time, where possible, to a reference standard run under similar conditions.

Done in that spirit, a chromatogram stops being three magic letters on a coversheet and starts being what it always was: a careful, time-stamped, signed photograph of a separation. It will not tell you everything. But it tells you the thing it tells you with unusual honesty, and that is rarer than it should be.

Sources

• Mikhail Tsvet — Wikipedia. Biography and history of his 1900–1906 chromatography work.
• Encyclopaedia Britannica — Mikhail Semyonovich Tsvet, biography and contribution to chromatography.
• Csaba Horváth — Wikipedia. Hungarian-American engineer credited with building the first modern HPLC.
• LCGC International — “Csaba Horváth and the Development of the First Modern High Performance Liquid Chromatograph.”
• United States Pharmacopeia, General Chapter <621> Chromatography (revised 2022).
• Royal Society of Chemistry — Chromatography teaching resource (RSC Education).
• Royal Society of Chemistry, Analytical Methods Committee — “Guide to Achieving Reliable Quantitative LC-MS Measurements.”
• Pitt, J.J. (2009). “Principles and Applications of Liquid Chromatography-Mass Spectrometry in Clinical Biochemistry.” Clinical Biochemist Reviews, 30(1): 19–34.
• ScienceDirect — Ultra-High-Performance Liquid Chromatography: technical overview of sub-2-µm particle columns and high-pressure operation.