first try to understand where does contamination come from.

my suggestions:

mobile phase - simple to check

Initial mobile phase  - simple to check

sample preparations (often drying step affects on presence of contamination)

column - longer washing/equilibrating step after each injection? 

i usually use washing step after probable high concentration samples and before low concentration samples.

try to change capillary, wash ESI compartment (orifice, plate) with methanol 

hope it will help somehow

Remove column & run with union and see is it from column or system. If not from system then it is from column & flush the column with 100% Acetonitrile for 1hr at high flow rate. And if it is from system change the needle wash solvent like 10:40:50(water:Methanol:acetonitrile) and increase the needle wash time. And if you source is contaminated then clean the cone/curtain plate and flush the probe's capillary at high rate with methanol. And also check for any contamination in mobile phase or additives.

HI,

As it was mentioned you have to locate the problem:

-If initial blank sample is affected check the sample preparation steps,

-Check the MS 

-Check the LC

-Check the tubing 

What is the LC manufacturer's carry over specification?  You may be fighting against something that you cannot fix.  You can increase needle wash volume.  You can increase a column equilibration time (a good measure is that if it takes 5 min for analytes to elute from column, then it will take the same or longer time to wash out the column and equilibrate.  Another thing that you can do is to arrange a sequence of running samples.  Start with ascending order in concentration, and then descending order.  Run a blank-blank (just eluting solution mix) after the last descending sample to show that there is a minimum carry over, etc.

Hi

How does the carry-over look like in your run? Is it a constant signal which varies over time or is it a clear peak in your blank injection with the same retention time? I suppose you are running a gradient, did you ever try to run your gradient in a loop? I mean you just repeat your gradient composition several times in the same run. As long as you observe an aditional peak in your second gradient procedure it means that your carry-over is caused by column (and tubes) and maybe by your eluents.  In this case you have to change your eluent and to add more time for flushing the column at the end of your run. If you only observe one signal in your blank then it must be the injection system. In this case you have to adjust the washing procedure of your injection system. If its a more or less constant background signal it is caues by your ion-source, which needs cleaning.

First, you can check if the unexpected peak is resulted from LC water, solvent or column used by injecting control blank samples. Then examine if this contamination can be reduced by repeating needle wash.

We had the same problem with our Agilent HPLC 1260 and 1100 with ciclosporin. The needle wash is not so effective like with Waters HPLC's. About 0.3% of the sample concentration we could find in the next chromatogram. We increase the needle wash volume and make blind samples between two real samples.

A not sufficient needle wash may be possible. We had a similiar problem with an Agilent system. But we found a worned rotor of the Rheodyne valve as reason.

Thank you guys, as per my discussion with the expert in USDA, who worked in the compound "Tulathromycin"; the carry over due to the nature/chemistry of the compound. As advised by him, we washed the syringe by using 50% formic acid (Wash A) and Methanol (Wash B) pre and post injection. We are injecting blanks between the samples as well. We are quite comfortable and confident with our results. The carry over problem reduced. Thank you all. 

You can Use Magic Mixture for washing the injection needle. 

Magic mixture : H2O: ACN: MeOH: Isopropanol (25: 25: 25: 25 V/V)

This article may provide you or anyone else with some ideas on how to diagnose and prevent carry-over issues in HPLC and LC/MS.

"Carry-Over (Carryover) Contamination in HPLC and LC-MS Systems"; http://hplctips.blogspot.com/2015/02/carry-over-carryover-contamination-in.html

Additionally, improper use of the needle wash feature can result in carry-over problems too (e.g. If you have a modern Agilent A/S, please do not use the needle wash feature as it is not needed for 99% of samples due to the A/I design. Be sure to obtain training in how your A/I works before use. Agilent uses a very different design with a FLOW through system and this is completely different than the types used by Waters, Shimadzu etc). Poor method development, a fouled column and of course hardware problems (e.g. worn injector rotary valve seal) are also major contributors to the problem. Proper system maintenance and training in how to use your system will solve the problem. The above article link will provide more information.

It might be due to long runner in sample or it may be ghost peak which was appeared in next chromatogram.

Sample carryover is a significant problem that occurs in high-performance liquid chromatography (HPLC) analysis. Carryover effects cannot be tolerated in any high-performance liquid chromatography-mass spectroscopy (HPLC-MS) separation system, and proteomics analysis must be performed in a separation system with virtually no carryover. We have developed and optimized a cleaning method for eliminating carryover caused by the autosampler and the trap column. The new washing method uses an injection of trifluoroethanol into the injection path and onto the trap column to remove strongly bound peptides and proteins, and it includes trifluoroethanol as an additional solvent in the chromatographic mobile phase for enhanced cleaning of the separation column. By application of this method, a significant reduction in carryover was achieved without any loss in the amount of proteins and peptides identified by MS.

how to avoid metformin carryover in HPLC method