HPLC Products

A “no peaks” situation can be very frustrating, especially if everything appears to be as it should be. My recommendation is to assume nothing, and work things though systematically. This list is by no means definitive, but might help to highlight potential causes.  First define the problem:

• Is there flow coming out of the waste pipe?
• At the right rate?
• System backpressure OK?
• Are there no peaks, or just not in the right places?
• How flat is the baseline? If it is absolutely flat, don’t look further than the data system! If its flat but you see system noise, start with the detector. If it is normal, but the peaks just don’t come out, start at the beginning and work through!

Problems with the Eluent

Do you know for certain what’s in the/each eluent bottle? Is it fresh?

Wrong %B             The composition of the eluent might be wrong. Too weak and all the peaks can stay on the column, too strong and they all come out with the solvent front.

Wrong pH              This is unlikely to cause no peaks, but might cause some peaks to co-elute

Wrong Solvent    If the eluent strength is different, the %B situation applies. But it might be that the HPLC fairies have topped up your acetonitrile bottle with acetone instead.  (Even suppliers can make this mistake sometimes, so check the smell of the eluent to be sure it is what it says on the tin!)

Wrong UV Cut-off. All solvents have a UV cut-off (the wavelength below which they absorb all UV light). If you pick (eg) normal acetonitrile instead of Far UV, and try to work at 205nm, you will find out what I mean. Other examples are using acetic acid below 250nm, using stabilised THF instead of unstabilised (the stabiliser, BHT, is immiscible with water and absorbs UV at all wavelengths below about 300nm), or using normal phase eluents such as ethyl acetate or acetone below about 300nm.

Buffer Concentration. This is unlikely to cause no peaks, but if the buffer concentration is too high, all peaks may elute with the solvent front. Too low and the buffer will be ineffective.

Wrong Channels selected on a gradient pump. Its an easy mistake to make, but if the solvents required are on channels B and C, and the gradient is set to run A and B, you may get no peaks.

Problems with the Sample and Solvent

Wrong Sample.    In a busy lab it is possible to select the wrong sample, or even the wrong vial position in an autosampler. Sometimes a blank is injected (the only time when no peaks is a refreshing change!)

Wrong dilution.   Maybe 1 in 100 instead of 1 in 10. Maybe below the detection limit of the detector.

Wrong sample solvent. If the sample solvent is a very strong eluent it can cause the peaks to elute very quickly. Ideally the sample solvent should be the eluent.  If the wrong sample solvent has been accidentally used, it might be that the sample is virtually insoluble, and hence no peaks.

Problems with the Injection System

1                Manual Valve

Injecting in the Inject Position instead of the Load position. We have probably all done it! And the port labels don’t help. But for a Rheodyne valve you inject in the Load position and rotate the valve clockwise!

Waste pipe exit not at the same level as the inject port. If you have the waste pipes ending much lower than the valve, a siphon action can draw the sample through the loop and into the waste

Waste pipe exit blocked. It is important to flush the valve thoroughly to prevent sample crossover, but if the sample is fairly concentrated, or consists of a solid material dissolved in a solvent, it is likely to dry in the exit of the waste pipe. When this happens, it becomes difficult or impossible to inject samples, and the new sample is forced out through the injection port. No sample enters the loop, and hence no peaks

Wrong loop size. Too small and the peaks might be too small to see

Valve faulty. If the valve is leaking, the injected sample may leak out rather than pass through the loop. If this happens, no waste will be observed leaving the waste pipe, and there will be no peaks. The valve will need overhauling.

2                Autosampler

Wrong sample vial range. Its just a programming error, but it fools the autosampler!

Needle bent. If the x-y calibration of the autosampler goes out of alignment, the need will hit something solid like a vial cap and bend. So if this happens, not only is no sample collected, but you need a new needle, and the x-y calibration must be reset to prevent it happening again.

Needle blocked. This happens when the needle picks up a core of rubber from the septum. It can sometimes be cleared with needle wire, or possibly compressed air (very carefully!!) If it happens too often, consider using different septa. Call us if you are not sure what to use.

Air in the syringe. Air does not usually stay in the syringe for long, but if it gets injected, you may not get much in the way of peaks. Air comes from bubbles in the wash solution, or if the wash solution has run out, or if there is not enough sample in the vial.

Autosampler inject out of synchronisation with system start. If the autosampler injection is not synchronised with the gradient and data system starts, you may not see peaks, or at least not where you expect them! If the autosampler is controlling the starts, make sure it has the longest run time, so that everything else is ready when the next inject signal comes.

Needle depth too high. This is adjustable in an autosampler method, to allow for the use of different types of vials. Too high, and you can’t use all the sample, too low, and it drives a hole through the bottom of the vial.

Too little sample. You need at least the injection volume x the number of injections + the minimum vial level for the needle. Otherwise it sucks air!

Too small injection volume. Could have chosen 2ul instead of 20ul etc

Problems with the Pumping System

• Is the pump running?
• Is solvent dripping out of the waste pipe?
• Is it dripping at the correct rate?
• Backpressure normal?

Flow not started. On some pumps you set the flow rate separately from the start/stop button. Flow may not have been started. This can also happen if you set a low-pressure cut-out before starting the pump. Unless you hold the start button until the pressure exceeds the low pressure cut-out, the pump will stop

Flow stopped – Overpressure. If something has caused the pressure to exceed the cut-out level, check that the level is set at a sensible level. (eg if you are used to working in bars, and you set the limit to 400, a pump working in psi will stop quite quickly!). If it is a genuine overpressure situation, this must be addressed before flow can be restarted.

Wrong Flow Rate. If you’ve set 0.1ml/min instead of 1.0ml/min, peaks will take a while to elute! Backpressure and drop rate in the waste bottle give this away straightaway!

Leak. A leak from the piston seal will result in low flow rate, a low system back pressure, and a low drop rate in the waste bottle. It will usually result in a puddle under the front of the pump, unless you use piston backwash, when it leaks into the backwash solution so you don’t get the puddle! However if the level in the backwash bottle rises, this is a definite indication of a leaking seal.

Air Bubbles. Air in the pump is a very common fault. A good degasser makes the problem go away. But if air does form in the pump, the flow will be greatly reduced and hence retention times will be much longer, and it may appear that there are no peaks. The give-away is a fluctuating and low pressure readout.

Check Valves faulty. Leaking check valves mean that the actual flow is less than the selected flow. Hence the pressure is low (and sometimes fluctuating), the retention times are longer, and the drop rate into the waste bottle is slow.

Wrong Gradient Run Time. If the gradient time is set too long, peaks will come out much more slowly

Problems with the Column

Wrong Column.   Check in the column oven that you have the right column connected! If it is the wrong type of column, you may well get no peaks, or no retention.

Bonded Phase stripped off. If the operating pH is below pH2, it is likely that the bonded phase (C18 or whatever) has been stripped off the column, reducing it to a silica column. This is usually a gradual change, and means that the column will change to acting like a normal phase column. Polar materials may then be irreversibly retained.

Wrong column temperature. If the column oven is off, retention times will be much longer. Back pressure will be higher, but the solvent front time will be the same.

Inlet tube to column slipped out. No sample of flow will pass through the column, and hence no peaks. However it should be obvious, because the back-pressure will be virtually zero and no solvent will drip into the waste. Inside a column oven though, a leak is not easily seen.

Blocked frit or column. Will give such a high back-pressure that the pump cuts out.

Problems with the Detector

To start with, check the detector is plugged in, switched on, and the lights come on. At least then you know there are no fuses blown etc

Detector lamp blown.  If the detector lamp blows, the baseline is likely to be all over the place, as it subtracts nothing from nothing without much success. You will get no peaks but BIG noise. The ‘lamp on’ light will not be on, the lamp energy level (if available) will be zero, there will be no pale blue glow from around the lamp housing.   (Always wear UV safety glasses if you want to look inside).

Wrong wavelength or voltage. For a UV detector this is the wavelength, for an electrochemical detector it’s the voltage. Too high a wavelength or too low a voltage and you won’t detect the peaks as they come out.

Wrong range setting. Many people do not use the range setting. But if you have very small peaks and do use it, if it is then reset to 2AUFS you will see next to nothing.

Output voltage incompatible with data system. Most data systems require a 1V unranged input. So if you connect to the 10mV output from a detector, at best you will get very small peaks!

Data system cable not correctly connected at the back. If using an analogue cable (ie 1V unranged instead of RS232), check that it is securely connected with no broken wires. If it is disconnected, the cable will induce enough electric signals to give noise, but of course, no peaks.

Flowcell cracked or contaminated. Too high a back-pressure across the cell will cause the quartz windows to crack. It can be hard to see, but much of the light beam is obscured. Hence the response to peaks will be very small. It will probably give a very high starting signal, which can of course be called zero by pressing Autozero, but then you get no peaks! There may also be a leak from around the flow cell housing, and if so, the system back pressure will be normal, but the drop rate into the waste bottle will be low. Contamination gives a similar result without the leak. Be careful about cleaning the cell. Draw wash solution through the cell by sucking with a syringe. Squeezing it through with positive pressure can crack the windows very easily.

Air in the cell. Air which is held in solution by high system back-pressure may be released when the pressure drops at the outlet to the column, and ends up in the flow cell. The effect is the same as a contaminated cell, except that the noise is much worse because the bubble won’t sit still. To get rid of a bubble, block the waste pipe with a rubber septum for about 2 seconds at 1 ml/min flow and release suddenly. This causes the bubble to compress and position over the cell outlet, and when released it should come out of the pipe. Peaks are generated when air is in the cell, but they are very hard to see because of the massive noise that can be caused by the bubble.

Signal offscale. If the detector is turned on with solvent in the cell, the signal should be close to zero. If it is not, something (contaminated cell, material eluting from the column, eluent which absorbs UV light) is blocking the signal. If it goes offscale, no further peaks will be detected. You may be able to call the signal zero by pressing Autozero, but there will be little or no capacity to detect peaks.

Autozero being held. It is not uncommon to use a contact closure to zero the detector when an injection is made. This is done by connecting the ‘Autozero’ and ‘ground’ connectors to the start signal. However if you do this with a Rheodyne valve, it holds the start signal contact closure, thus effectively holding the Autozero command, and keeping the detector signal at zero. This can also happen if there is a fault with another device which incorporates peak detection (solvent recycler, fraction collector) and which is connected to the detector analogue data signal.

Cell or lamp cover not closed properly. Some detectors (eg Varian UV, ABI Fluorescence) have a door to the cell which must be securely closed or it  allows no signal. In the case of Varian, it turns off the lamp! So if the door is not securely closed, you will definitely get no peaks!

Problems with the Data System, A-D Converter, or Computer

A-D Converter not turned on. This applies only to systems such as Perkin Elmer Nelson, where the A-D converter is a box standing next to the computer. If the A-D system is a card inside the computer, then as long as it was correctly installed and set up, it will be active as soon as the computer is turned on a data system loaded

Incorrect Voltage setting. The voltage setting in the software must be the same as the analogue signal output from the detector. Otherwise if the data system is looking for a 10v signal, even fullscale 1V will only give a small peak.

Incorrect data Channel.   For data systems with more than one detector channel, make sure that the correct channel is being monitored. For systems with the facility for more than one HPLC system, make sure that the correct system is being monitored.

Data Collection not started. If the HPLC system is running but the data system did not start collecting, no peaks will be seen. Check the start signal is connected, and if it is a contact closure, check that all the positives and the negatives are connected the right way round! All devices detecting a contact closure apply a 4V DC signal, and when the voltage drops and a current flows, they use that as the start signal. SO if a + and – are connected the wrong way round, current will flow all the time, and the start signal will not be detected. Suspect this if you have just moved and reconnected an HPLC system.

Software hung up. This can sometimes happen, where the computer appears at a glance to be working, but upon closer inspection, has hung up. Suspect interrupt problems and look for what happens just before the crash. Eg a clash with the printer, the mouse etc. However before you call anyone any names, make sure that you haven't maximised the Data Acquisition window and then set it to 'Always on Top' . In this situation it could be waiting for a response from you, but the dialog window is hidden under your maximised data acquisition window, which remains Always on Top because you asked it to do so!