Hello Jairo: We need more information and some context to assist you solve this problem. Please provide the actual detailed method and conditions (e.g. flow rate, % composition, gradient or isocratic, inj volume, wash solvents, sample type, method). All I can see from the chromatogram you included is that you have a negative drifting baseline with negative peaks appearing at several time points. Hundreds of things could cause that including: improper detector or pump settings, failure to equilibrate or wash the column before use (most common reason is fouled column from incorrect wash procedure). Is this an analysis run or just a snapshot of the baseline in isocratic mode with no injection? What am I looking at? If you can provide more details, we may be able to narrow down some of the most likely causes.

BTW: I see that you have the bandwidth for your wavelength set at 1 nm. Please do not run any HPLC analysis with a bandwidth of 1nm. A value of 8 or 10nm is recommended for most types of analysis. Using a narrow bandwidth of "1" means you will gather a very tiny amount of data that is super selective and prone to high noise levels. Please refer to this article for more info:

• "UV / VIS, VWD, DAD, PDA HPLC DETECTOR SIGNAL BANDWIDTH (bw)"  http://hplctips.blogspot.com/2011/09/uv-vis-hplc-detector-signal-bandwidth.html

I also see that you have the Reference Wavelength turned 'OFF'. That is excellent! Always operate with the Reference Wavelength 'OFF'.

Some articles/links which may help you:

• "Common Causes of Baseline Noise in HPLC, UHPLC";  http://hplctips.blogspot.com/2014/09/common-causes-of-baseline-noise.html
• "HPLC Retention Time Drift, Change, Variability or Poor Reproducibility. Common Reasons for it."  http://hplctips.blogspot.com/2015/11/hplc-retention-time-drift-change.html
• "Diagnosing & Troubleshooting HPLC Pressure Fluctuation Problems (Unstable Baseline) ";  http://hplctips.blogspot.com/2014/01/diagnosing-troubleshooting-hplc.html

"

Since baseline at others wavelengths are nearly flat, but others are showing intense dips, it seems that something strongly adsorbed is leaching from the column, but why it is negative, I am not sure. Use another column, and see if it happens again.

Additionally, your data collection rate of 5 Hz is too low for any decent separation. Choose about 100 or 40 Hz. By choosing a response time of 2 s, you are compromising the column performance. We have seen efficiency drop from 20,000 plate to 4000 plates just by changing the response time of 2 s. Those things need to be addressed too, although this setting is not going to cause baseline drift.

Noise and poor integration quality can be caused by too long or short of a data collection rate. We would need to know what the flow rate and peak width of the narrowest peak in your sample is to determine the most appropriate data sampling rate (info which was not provided).

While UV/VIS detector sampling rate is not the issue in this case, you can read more about how to set it up correctly at this link:

• "Tip# 112: Determining the Data Acquisition Rate (Sampling Rate) For Your HPLC Detector"; http://lc-ms.com/Tip_112_Determining_the_Data_Acquisition_Rate_For_Your_HPLC_Detector.htm

Hello Bill letter,

Thank you so much for your kindly answer and links suggestions. Your links are so useful to me to understand the problem with the baseline. The attached document had a snapshot of an isocratic system with ACN with no injection (figure 1) but using a different solvent system I had a similar baseline drift (negative peaks) during a sample analysis (figure 2). The eluent was a gradient mixture of A and D, where A is H2O-AcOH (991.7:8.3; v/v) and D is ACN-AcOH (992:8; v/v), the flow rate was 0,450 mL/min, the column was a C18 (Acclaim 120 A, 3 um, 2.1x100 mm). The injection volume was 5 uL, and the sample was a flavonoid type. I read yor links and one of them suggest a change of temperature, I shut down the air conditioner and run another sample analysis at room temperature (25 °C) (figure 3), the other experiments was at room temperature with the air conditioner at 18 °C, and the negative peak disappear but still drifting, how I can resolve it?

Hello M. Farooq Wahab,

Thank you so much for your kindly answer and suggestions for the data collection rate, in the next experiments I will be use it at 100 or 40 Hz at 2 seconds.

Best regards Jairo

Jairo: Please do not just "set" your sampling rate to a value of 100 or 40Hz without first calculating what it should be for your specific method. It should not be a general #, but a value based on actual sample data. The article linked in my previous post, "Tip# 112: Determining the Data Acquisition Rate (Sampling Rate) For Your HPLC Detector", provides a simple means to do this.

Thank you for providing a bit more info about your method. Unfortunately, you still have not provided the %composition profile for your gradient run.

Looking at the two chromatograms you have provided, both of isocratic runs with different liquids. They both appear to show the same thing. That is, elution of "something" in the 300nm region after some time period.  This is secondary to drift. Drift is another matter all together and its causes are too many for this site to handle. Now, the negative peaks you are see'ing look "real" so I will again suggest that your column has not been washed after analysis and continues to bleed some sample material over time. The solution to the problem is to incorporate a strong solvent into the wash phase, after analysis, but also as part of your HPLC gradient method. A good method will run a wide range of solvent properties to elute EVERYTHING off during the run. You are using only ACN. Perhaps your samples are not fully soluble in this and you need to use some methanol (which has better properties), or something else. IOW: Develop a better method to begin with, then develop a wash method to re-equilibrate the column. Once that is done, nothing should elute from the column during equilibration. One more note: You may have contaminated your column (fouled the frit or column head) so if possible, replace the column with a new one and see if you have the same problem. If it goes away, then proceed to develop a better method so this does not happen again.

Hello Jairu, Although 40 Hz might be more than enough for your work, choosing a response time of 2.00 s is unrealistically slow. Try 40 Hz with a response time of < 0.2 s and see how the chromatograms look like. Also see how the above mentioned calculation helps you. You will see a magical improvement in peak shape when you reduce the response time. We have written an extremely detailed theoretical and practical review on sampling frequency and response time and it will appear online soon in Analytica Chimica Acta. What I want to make clear is that there is no theoretical basis of choosing a sampling frequency as long as it is above the limits of sampling theorem which usually suggests a frequency of 2-3 Hz!! In reality, this is not at all enough. The work is being delayed because the Elsevier publishing offices are flooded in India.

You can try to invert your column, just change its direction. Sometimes the problem appears in the beginning of the column and when you turn it, the problem (might be a physical crack) will be at the bottom of the stream, and will not influence your results. Worked for me for several times.

Dear M. Farooq Wahab,

I really appreciate your opinion and yours observation for this problem. I would like to read your review with the explanations to have more information about HPLC works. I will do a new analysis using your observations along with above mentioned calculation and I will see what happen, results will communicate by this way.

In your experience, What do you think about the use inverted column to wash it? suggested by Helen Domeshek in a previous comment.

Jairo: Reversing the column may help, but it will not correct for the main problem. You must improve your HPLC method first. Cleaning the column you have is a temporary fix until you change your method to incorporate a stronger solvent which retains, then elutes ALL of the material during each analysis. *Improve the method and techniques used.

Dear Bill Letter,

the detailed method that I used to analyze my sample is described here.

peak width: 0.100 min

data collection rate: 5 Hz

response time: 2 seconds

maximum flow acceleration: 0.079 mL/min

maximum flow deceleration: 1.001 mL/min

pressure lower limit: 0 psi

pressure upper limit: 8992 psi

flow cell detection 13 uL

detection at: 360:1, 280:1, 330:1 and 260:1

3DFIELD. Min.Wavelength: 200:1

3DFIELD. Max.Wavelength: 800:1

3DFIELD. Ref.Wavelength: off

linnear multi-step gradient: solvent A: H2O-AcOH (991.7:8.3; v/v), solvent B: ACN-AcOH (992:8; v/v)

time (min) solvent A (%) solvent B (%)

  0                    88                 12

5.3                   88                 12

10                   79.5             20.5

21.3                 56                44

22                     0                 100

25.3                  0                 100

26                     88                12

32.6                 88                 12  

the period from 22-32 min was used to wash the column and restored the initial condition. Do you think that I should be use another solvent (like MeOH) to wash the column after sample analysis and before to restore the initial condition with solvent A and B?

Attached document had several printscreen for the detalied method

Dear Helen Domeshek,

Thank you so much for your opinion and yours observation for this problem. I will use it as a temporary fix before improve the method using the observation above described.

Sincerely yours, Jairo Bermúdez.

Jairo: Thank you for the additional info, but without an actual chromatogram showing your sample(s), it is very hard to know which parameters need to be adjusted. As such, I can only make very generalized comments.

(1) As noted earlier, your bandwidth settings of 1 nm really does not make any sense and I would strongly suggest you never use such small values. Try 8 or 10 nm to start with.

(2) The chosen sampling rate of 5Hz is probably too slow, but as noted before, it should be calculated based on your actual sample peaks (not shown), not randomly assigned. *Sampling rate has nothing to do with the extra negative peaks you see. Use the information provided to calculate the correct sampling rate for each method. * Training is the key to learning.

(3) Your analysis method ends at 21.3 minutes where the concentration of ACN is only 44%. You then quickly change to 100% ACN at 22 minutes to begin to wash the column. Maybe this makes sense, but we have no chromatograms to look at so do not know how this gradient profile relates to the proposed analysis. Allow me to make two suggestions. First, Do not create one method which runs the analysis AND washes the column at the same time. A far better technique is to use two different methods. One method for the sample analysis and one method for the wash and equilibration steps. Your wash method should be aggressive and separating it away from the analysis method keeps your analysis chromatogram nice and neat so you do not have to worry about integrating all of those unrelated peaks. It also reduces the total analysis time too. Next, how do you know that all of the peaks have eluted from the column at 21.3 minutes? To be sure, you should first change the gradient program to incorporate a hold period or step with a much high concentration of ACN (perhaps to 95%, never 100%). Then "HOLD" that concentration for a period of time until the baseline is flat and no peaks elute. You also need to confirm that if you change to a better, but slightly weaker solvent such as MeOH, that no additional peaks are seen. This can only be done by experimentation and is part of the normal method development process (trying many different methods to resolve ALL peaks), 

Is there someone local that can help you with these HPLC fundamentals? You will learn much faster with someone there to teach you method development. Troubleshooting these types of issues is easy for an experienced person to do on site, but hard to do via the Internet. Chromatography is a complex technique to learn and use.