Common Causes of Baseline Noise in HPLC, UHPLC.

Achieving a flat baseline which does not exhibit spikes, ghost peaks, drift or wander in an unpredictable manner should be a primary goal when performing HPLC analysis or developing methods. Methods which result in flat baselines and have well defined, sharp peaks allow for accurate sample area integration. Integration algorithms perform poorly in quantifying peaks on sloped, drifting or noisy baselines. Excessive baseline noise contributes to many problems, including poor quantitation, high %RSD errors, peak identification errors, retention time variation and many other critical problems. Properly developed HPLC methods are reproducible methods which apply and utilize good chromatography fundamentals. Note: "Noise" is a relative term, often w/o meaning. You should always describe it scientifically, measure and compare the signal to noise ration (S/N) of the baseline vs the peak plus note any cyclical patterns (useful in troubleshooting).

Note: A lack of proper training in the operation of the HPLC system, improper start-up or poor quality maintenance of the chromatograph (Examples: failure to degas and purge the system lines before use; poor mixing; an air bubble stuck in a check valve, a bad detector lamp or a leak will often result in baseline noise) are the main causes of noise. Your HPLC system must be optimized for your specific application. Be sure and allow time for the mobile phase to reach full equilibration with the system before starting any analysis. Do not start an analysis until the baseline is stable.

In this article, we will discuss how temperature fluctuations, inadequate mixing, inadequate degassing and flow cell contamination can result in excessive baseline noise. We will provide suggestions on how to reduce or eliminate these problems. Troubleshooting should be done on-site, not over the web or telephone.

TEMPERATURE FLUCTUATIONS:

To obtain reproducible results, the temperature of the HPLC column must be kept constant during each analysis. Laboratory room temperatures often vary by several degrees during the course of one day and these changes will often change the retention characteristics of the sample(s). The 'On' and 'Off' cycling of power from an air conditioner or heating unit will often cause the baseline to drift in a cyclical manner, up and down, during the day (this can often be seen as a clear sine wave pattern when you zoom-in to study the baseline trace over time). Temperature also changes the refractive index of the mobile phase. Light based detectors (UV/VIS, RI...) will show this change as drift up or down). In some cases, a temperature change of plus or minus one degree C from run-to-run can cause changes in retention times which effect reliability of the method. 

To reduce temperature fluctuations, you must control the temperature of the column and mobile phase (if applicable) during the analysis. This is most commonly done by: (a) using equilibrated mobile phase at the start of the day or analysis, (b) keeping the interconnecting lines as short as possible (esp. any which exit the column and go to detectors/flow cells), (c) insulating any stainless steel lines with plastic tubing to reduce heat loss and (d) using a thermostatted column compartment to maintain the column at a single set temperature throughout the day. Control of the column temperature will remove 'temperature' as a variable from your analysis. Temperature should be a constant run to run, not a variable. Be sure and document the temperature selected as part of your method.

INADEQUATE MOBILE PHASE MIXING:

The associated noise and ripple of incomplete mixing can reduce the limit of detection (LOD) and increase integration error. Both high pressure (with separate pumps) and low pressure pumping (one pump with a multi-channel proportioning valve) systems depend on efficient mixing to reduce noise. For gradient analysis, failure to completely mix the mobile phase solution before it enters the HPLC column often results in excessive baseline noise, spikes and poor reproducibility. "Mixing" is often initially accomplished by combining the flow paths of more than one solvent channel together, using a multi-channel gradient valve or tubing. Mixing also performed directly in a mixer installed in the flow path of an HPLC pump. This mixer is often a static mixer (a simple 'Tee', a tube filled with baffles, a frit or beads, valve orifice or microfluidic device) of low volume design for chromatography use, but allows adequate mixing of the liquids within a prescribed flow rate range. The best mixers incorporate longitudinal and radial mixing in-line. A mixer with too low a volume or of insufficient design can result in poor mixing of the mobile phase (note: incorrect solvent compressibility settings can also cause mixing and noise problems too). To reduce mixing problems, first insure that the mobile phases used are fully soluble with each other. Next, make sure that any mixer used is appropriate for the flow rates and volumes you will be using. If needed, run a gradient valve test to insure that each valve channel is working properly, not leaking or introducing any cross-flow leakage to another channel. Monitor the baseline for pressure stability (% ripple), drift and artifacts  (e.g. spikes) in real time to spot problems and make adjustments to correct them. 

INADEQUATE MOBILE PHASE DEGASSING:

For the best results, continuously degas your mobile phase. Reducing the amount of gas will also improve signal to noise levels of detection, reduce drift and reduce pump cavitation. If you are using an electronic vacuum degassing module, make sure it is maintained and working 100%. A faulty degasser may cause more damage (contamination) to your system and methods. Maintain and Repair them just as you do for your other instrument modules. Gas bubbles may cause check valves to malfunction (get stuck), baseline noise spikes to appear randomly, flow rates and/or pressures to become irregular, detector outputs to show high levels of noise (from air in the flow cell) and also cause the loss of prime or cavitation in pumps. To achieve the best balance of low noise levels and high reliability, both aqueous and organic mobile phases should be fully degassed before and during use. This can be accomplished through stand-alone inline vacuum degassing modules or through gentle continuous helium gas sparging (*Helium makes an excellent choice of gas as it is not soluble in the mobile phase. Never use Nitrogen or Argon gas, they are soluble in the liquid!). In all cases, degassing must be continuous (not just done one time). Continuous degassing reduces cyclical noise and signal variations. For this reason, I do not recommend using ultrasonic baths to degas mobile phase solutions as these are not used in a continuous mode. The mobile phase solution starts to re-absorb gas as soon as you stop sonicating the solution. This results in continuous baseline drift (up and down).

Removal of gasses is critical to the function of a modern HPLC pumping system. The liquids used are compressed to very high levels which forces out solubilized gas from the solutions. This is best accomplished before the liquid is transferred into the pump. These gas bubbles must be minimized to achieve desirable baselines. *Even if you use a high pressure pumping system, an inline degassing system reduces the amount of noise and baseline drift. Properly maintain and service your degasser to insure compliant operation. IOW: Whichever method you use, always degas your mobile phase solutions.

FLOW CELLS:

One other less common cause of baseline spikes and random noise is due to either a dirty flow cell (i.e. the windows) or an air bubble trapped inside the flow cell. If the flow cell is suspected of having one of these problems, then it should be carefully rinsed or flushed out with an appropriate mixture of suitable solutions to expel the air bubble or remove the contamination. If possible, keep a spare, 'known good' flow cell on hand to swap out for troubleshooting purposes. This can help to quickly determine where the problem is. This flow cell must be the exact same size and type (volume and path length) for this purpose. If the cell's windows are contaminated and flushing does not restore them, then many manufacturer's offer kits which allow you to replace the windows and gaskets used. Warning: When attempting to clean or repair any flow cell, be sure and work within the manufacturer's operational specifications for the specific flow cell. Some flow cells are not designed to withstand even very low back pressure and damage can result if you exceed their maximum pressure or chemical rating.

Many other types of problems not mentioned in this short article can also cause baseline noise. For example, a sticking inlet or outlet valve on the pump, worn piston seals, worn out detector lamp(s) or detector electrode (EC) can induce noise. In all cases, the cause must be investigated in a logical, step-wise manner. Demonstrate what is working and rule out items one-by-one.

Reference: http://hplctips.blogspot.com/2014/01/diagnosing-troubleshooting-hplc.html

USP Guideline Note: HPLC Column Diameter Changes to Maintain Flow Rate Linear Velocity

USP Allowed Variations in HPLC Column Diameter (*USP 32, Second Supplement, Dec 1, 2009). In the previous USP version, a change of up to 50% of the flow rate was allowed. This has been changed in the more recent version. A wide range of column diameter changes are now allowed, provided that the linear velocity is kept constant. *We addressed the effect of changing column diameter on flow rate in a previous blog post, but this time I have also expanded on the calculation by including the extra variable for column length (L1 and L2) as well.

*Adjusting the Column Flow Rate for Changes in the HPLC Column Diameter.

Linear Velocity Formula:

   New Linear Flow Rate₂ = Flow Rate₁ x (L2 x D2²) / (L1 x D1²)

Flow Rates are in ml/min.

L1 = Column Length (original) in mm.

L2 = Column Length (proposed) in mm.

D1 = Column Diameter (original) in mm.

D2 = Column Diameter (proposed) in mm.

Example #1:

Original column is a 150mm x 4.6mm (L x ID) used at 1.000 ml min. We would like to determine what the equivalent flow rate (F₂) would be for a column which is 150mm x 2.1mm (L x ID) to maintain the same linear velocity. This is a proposed change in column diameter of > 50% so it would not have been allowed under the old guidelines. The newer guidelines take into account that with the same particle size, changing the column diameter will not change the chromatography if the linear velocity is maintained as before. Let’s calculate the new flow rate using the formula above.

1.000 x (150 x 2.1²) / (150 x 4.6²) = F₂

                    1 x (661.50 / 3,174) = F₂

                              0.208 ml/min = F₂

Popular LC/MS and HPLC Volatile Mobile Phase Modifiers

For applications which utilize an Evaporative Light Scattering Detector (E.L.S.D.), Charged Aerosol Detector (CAD) and/or Mass Spectrometer Detector with Electrospray Ionization source (e.g. LC/MS, MSD or LC/MS/MS), a fully volatile buffering system is usually required. Many of the common HPLC buffers such as sodium or potassium phosphate are not compatible.Use the smallest amount of buffer which provides buffering under the analysis conditions (e.g. 10mM). *Select a buffering agent (or modifier) which are within 2 pH units (+/- 1) of the sample's pKa and 2 pH units away from any acid's pKa. 

• For LC/MS applications: Positive ion mode favors acidic mobile phases and Negative ion mode favors basic mobile phases. However, feel free to experiment using both ionization modes and don't forget about using adducts (e.g. ammonium and sodium) with all types of samples to improve signal response. *Maintain these buffers at or below 10 mM. Adjust the pH of the mobile phase to be 1 to 2 units away from your sample's pKa.

Table 1:  Popular examples of useful volatile mobile phase buffers, modifiers and/or additives.

BUFFERING/MODIFIER AGENT                                   USEFUL pH RANGE

• Ammonium formate                                 2.8 - 4.8; 8.2. - 10.2
• Formic Acid                                            3.3 - 4.3
• Pyridine/Formic Acid                               3.3. 4.3, 4.8 - 5.8
• Trimethylamine/Formic Acid                     3.3 - 4.3, 9.3 - 10.3
• Ammonium Acetate                                  3.8 - 5.8; 8.2 - 10.2
• Acetic Acid                                              4.3 - 5.3
• Trimethylamine/Acetic Acid                      4.3 -5.3, 9.3 - 10.3
• Ammonia/Formic Acid                              3.3 - 4.3, 8.8 - 9.8
• Ammonia/Acetic Acid                               4.3 - 5.3, 8.8 - 9.8
• Ammonium Bicarbonate                           5.9 - 6.9,  8.8 - 9.8
• Ammonium Carbonate                              5.9 - 6.9, 8.8 - 9.8  
• Carbonic Acid                                            6 - 8 (pKa 6.37/pKb 7.63)
• 1-Methylpiperidene                                   10.0 - 12.0  

• Trifluoroacetic Acid (TFA)                        pKa = 0.3 (WARNING when used with MS

                                                                                     systems!).  See notes #2 and #4 below.           

*Notes: (1) Formic acid (3.75) is slightly stronger and more volatile than Acetic acid (4.75). Formic acid is often available in higher purity grades and absorbs less in the UV region making it a better choice for most chromatography applications. It works well in positive mode LC/MS analysis, esp at 0.1%. (2) Trifluoroacetic acid (TFA, pKa = 0.3) is very strong and volatile, but we do not recommend its use in LC/MS applications as it can increase the background signal levels (esp. in Negative Mode) LC/MS (m/z 113), be very hard to remove from the source and result in long term instrument contamination. Difluoroacetic acid (DFA) and ammonium formate are other alternatives as they offer good ion pairing capacity with less ion suppression problems. (3) Triethylamine (TEA, pKa 11) is volatile, strong and very stable, but causes similar contamination problems resulting in high background signals when used in Positive Mode LC/MS (m/z 102). (4) Many ion-pairing reagents suppress ionization, bind to the plastics and metals used and contaminate the flow path. If you must use them, please do so using the lowest possible concentrations levels and thoroughly decontaminate the entire flow path of the system after use (or dedicate the MS system to use with them only). Minimize further contamination by labeling and using a dedicated column for the application (Do not use that same column exposed to ion pairing compounds for any other methods or applications). (5) Acids and bases alone provide little "buffering" so should be used with a secondary buffering species to resist change in pH.

Gradient Mixing Test For Your HPLC Pump (Step Gradient)

The most popular type of gradient pumping module used to perform HPLC analysis utilizes a low pressure mixing valve in their design. These valves are electronically controlled and proportion the amount of mobile phase from one of several solvent channels into a mixer for introduction to the pump head (*the solenoid valves used for this are sometimes called gradient proportioning valves). They provide random access to multiple solvents (e.g. 4) for method development and column flushing. The mobile phase solutions are mixed at low pressure before entering the high pressure side of the pump head (where they undergo compression). This design requires only one high pressure pumping head and can allow for very high mixing accuracy (often 0.1% per channel) of the mobile phase. This allows for the formation of mobile phase gradients over time which greatly aid in resolving samples apart on the column.

The gradient proportioning valves need to be tested along with the other parts of your HPLC system on a regular basis to insure they are operating within the manufacturer's specifications. They should also be tested anytime you suspect a problem may be present. One quick way to check the operation of two of the valves is to use a tracer compound and STEP gradient to monitor their operation. You can set up a method to perform this test as suggested below.

QUICK GRADIENT COMPOSITION TEST:

Bottle A = 100% DH20;
Bottle B = 0.1 % Acetone in DH20 (*Acetone is the tracer compound);

Flow Rate = 1.000 ml/min;
Column = No column. Install a restriction capillary in place of the column to obtain a backpressure of > 60 Bars;

Detection = 265nm (10 nm bandwidth) UV;

STEP Gradient Program:
    0 to 2.00 min, 0 % B
    2.01 min, 20% B
    4.01 min, 40% B
    6.01 min, 60% B
    8.01 min, 80% B
  10.01 min 100 % B
  12.01 min 20% B
  14.00 min 20% B

Note: If the delay volume (dwell volume) of your system is large, then you may want to adjust the time values shown to LARGER values (i.e. 2 minutes delays are used in this example, but 5 or even 10 minute delays between steps may be more appropriate if your system has > 1 ml dwell volume.

Running the above method should result in a signal trace which shows a step-wise rise to 12.00 minutes (as the acetone concentration increases). The edges of the "steps" should be sharp and the risers should also be close to vertical. The final step change which starts at 10.01 minutes shows a linear gradient change back down to the 20% B level. This line should not have any bumps or dips in it and should transition smoothly back down. The height of the baseline at this point should match the height seen between 2.01 and 4.00 minutes (same 20% B). The height of the proportional steps (e.g. 20, 40, 60, 80) should also be the same. You can use your CDS to measure these height values.

Another useful aspect to view is the S/N ratio at each step. Use your CDS to establish noise windows within each range (e.g. 2.50 to 3.5 minutes). This data is useful when comparing the performance of the pump at different intervals.

If you observe deviations in the height of the proportional steps or dips in the lines, these can be caused by leaking or sticking check valves as well as leaking or sticking gradient proportioning valves. *If you have a quaternary pump, be sure and test all four of the valves used (2x per test).

Lastly, the above example is a generalized method and may or may not be applicable to your specific HPLC pump. Be sure and customize a test method which takes into account the pressure ranges, flow rates, delay volume, mixing volume, and number of low pressure channels used in your pump.