Polymerization of THF in THF/water mobile phase?
Posted: Sat Jul 30, 2005 5:46 am
Greetings all,
I hesitate to post this, risking being branded as a freak-fool, but I see that this rarely stops anyone else...
I've been doing some method development recently in GPC/SEC with PSDVB and silica columns and two-component mobile phases. Long story. The short of it is that when my mobile phase consists of THF and water, regardless of ratio and regardless of column type or detector (UV @234nm or RI), when I inject mobile phase (or make samples in mobile phase) I see a negative peak well resolved from the permeation limit that grows over time. In other words, 1) This negative peak shows up pretty early to be part of the solvent front. 2) The size of this negative peak is proportional to the amount of time that has lapsed since the mobile phase aliquot was pulled from the jug and placed in the vial. I've tried vials from different manufacturers, different types of caps and septa material, vials made from different types of glass, glass and plastic transfer pipettes for filling these blank vials... the only thing that correlates with the size of the peak is the amount of time the mobile phase spends in the vial. The age of the mobile phase before an aliquot is pulled and placed in a vial is irrelevant.
This growing negative peak is not observed when 100% THF is the mobile phase and blank.
The RT of the negative peak corresponds to that of an organic molecule with a molecular weight somewhere between 800 and 1300 AMU, estimated. Assuming all the usual GPC assumptions.
I'm running on Agilent 1100 HPLC systems. If I bypass the inline degasser, the solvent front vanishes, but not the negative peak in question.
I've been searching the literature and discovered that THF is a monomer used in the production of certain polymers, such as Spandex. Further, one MSDS I found for THF warns "Dangerous polymerization may occur." Better yet, there was a patent application in the EU for a novel industrial process for polymerizing THF in the presence of water with a zeolite catalyst - zeolites are minerals consisting of alumina and silica. From what I gather, all THF needs to start polymerizing is contact with a cation. The specific properties of the cation and its associated anion will determine the rate of polymerization and final chain length. Am I misled?
You see where I'm going. Has anyone else out there worked with THF/water mixtures in glass and seen evidence for some limited polymerization of the THF? I've repeatedly asked myself the question, "What's different about the vial and the jug?" I've explored many potential differences: extractables from the vials, types of glass, dissolved gases, rates of evaporation... and nothing seems to be related to this negative peak. Except: the ratio of liquid surface area in contact with glass versus the volume of liquid. That ratio is always greater in the vial. If glass initiates/catalyzes the polymerization of THF in the presence of water, that reaction will always go faster in the vial. Right?
One possible problem with this hypothesis is that the pTHF would have to be pretty monodisperse to give one nice peak in GPC. But my impression from the literature is that the polymerization of THF is an equilibrium process. Would the reaction conditions dictate the chain length, so that over time we just get more chains, not longer chains?
All comments and insights welcome. If this is an intrinsic property of THF/water/glass systems, then surely GPC/SEC chromatographers have seen it before.
Thanks in advance!
khimik
I hesitate to post this, risking being branded as a freak-fool, but I see that this rarely stops anyone else...
I've been doing some method development recently in GPC/SEC with PSDVB and silica columns and two-component mobile phases. Long story. The short of it is that when my mobile phase consists of THF and water, regardless of ratio and regardless of column type or detector (UV @234nm or RI), when I inject mobile phase (or make samples in mobile phase) I see a negative peak well resolved from the permeation limit that grows over time. In other words, 1) This negative peak shows up pretty early to be part of the solvent front. 2) The size of this negative peak is proportional to the amount of time that has lapsed since the mobile phase aliquot was pulled from the jug and placed in the vial. I've tried vials from different manufacturers, different types of caps and septa material, vials made from different types of glass, glass and plastic transfer pipettes for filling these blank vials... the only thing that correlates with the size of the peak is the amount of time the mobile phase spends in the vial. The age of the mobile phase before an aliquot is pulled and placed in a vial is irrelevant.
This growing negative peak is not observed when 100% THF is the mobile phase and blank.
The RT of the negative peak corresponds to that of an organic molecule with a molecular weight somewhere between 800 and 1300 AMU, estimated. Assuming all the usual GPC assumptions.
I'm running on Agilent 1100 HPLC systems. If I bypass the inline degasser, the solvent front vanishes, but not the negative peak in question.
I've been searching the literature and discovered that THF is a monomer used in the production of certain polymers, such as Spandex. Further, one MSDS I found for THF warns "Dangerous polymerization may occur." Better yet, there was a patent application in the EU for a novel industrial process for polymerizing THF in the presence of water with a zeolite catalyst - zeolites are minerals consisting of alumina and silica. From what I gather, all THF needs to start polymerizing is contact with a cation. The specific properties of the cation and its associated anion will determine the rate of polymerization and final chain length. Am I misled?
You see where I'm going. Has anyone else out there worked with THF/water mixtures in glass and seen evidence for some limited polymerization of the THF? I've repeatedly asked myself the question, "What's different about the vial and the jug?" I've explored many potential differences: extractables from the vials, types of glass, dissolved gases, rates of evaporation... and nothing seems to be related to this negative peak. Except: the ratio of liquid surface area in contact with glass versus the volume of liquid. That ratio is always greater in the vial. If glass initiates/catalyzes the polymerization of THF in the presence of water, that reaction will always go faster in the vial. Right?
One possible problem with this hypothesis is that the pTHF would have to be pretty monodisperse to give one nice peak in GPC. But my impression from the literature is that the polymerization of THF is an equilibrium process. Would the reaction conditions dictate the chain length, so that over time we just get more chains, not longer chains?
All comments and insights welcome. If this is an intrinsic property of THF/water/glass systems, then surely GPC/SEC chromatographers have seen it before.
Thanks in advance!
khimik