I need some help figuring out an identity from EI spectrum

[KM-USA]

I need some help figuring out an identity from EI spectrum of unknown components in polyethylene glycol samples. I have a homogolous series in some polyethylene glycol materials, this series has a similar but different mass spectrum and retention times than the actual polyethylene glycol peaks, whether injected as is or as trimethylsilyl derivatives. The unknown series also gets derivatized, but I'm not sure if it takes 2TMS like the PEG series, or just one TMS. I'm posting the chromatograms and mass spectra of some of the smaller MW components, maybe someone here is real good at determining structure from these and can help me. Four attachments, first two are derivatized. I think once we figure out one of what we have, then the rest will fall into place. The samples should contain only carbon, hydrogen, and oxygen, and ethylene oxide is reacted (conditions unknown) to form the PEG. Thanks in advance for any help.

[chemstation]

Hi,

1) Nice Chromatography !, short column, fast temperature ramp, hydrogen maybe ?

2) I use a MTBSTFA derivative, to help me determine number of hydroxyl groups in unknowns.

3) I have read in books, (never done it myself as I have access to a GCMS with Chemical Ionisation NH3),
that lowering the ionisation voltage can show the molecular weight ion, so you can calculate
neutral losses for structural elucidation and Isotope pattern calculations.

4) Retention Indices data, can also be used to calculate Molecular formula information based upon known chemicals.

kind regards
Alex

[KM-USA]

Nice Chromatography !, short column, fast temperature ramp, hydrogen maybe ?

Ha ! My boss had to through in his 2 cents worth that I had overloaded the column, the sample was residue extracted then taken up in a few drops of DMF, then mixed with BSTFA, then injected overnight. No, I did NOT dilute more and re-inject, no reason too, this is a qualitative investigation.

The column was a DB-5 HT designed for high temperature compounds. I zoomed off the chromatograms to 16 minutes or so, there are many larger peaks, like the repeating PEG series, and the unknown series.

Interesting is that there really is glycerin at like 4.97 minutes, maybe from some propylene oxide in the original ethylene oxide, or maybe just a bunch of side reaction products are formed. We've observed this unknown series in both polyethylene glycol samples from some suppliers and in fatty alcohol ethoxylate samples from some suppliers.

Some fatty alcohol ethoxylate samples contain trace PEG only, none of the unknown series, must be different manufacturing processes.

[Consumer Products Guy]

I'm afraid that figuring out identities from mass spectra is not my strong suit. However, I would also be interested in what is present if you can identify that. So please post any updates.

[MSCHemist]

You can email the GC/MS data to myself sschoenfeld[at]innovaflavors.com and I'd be happy to see if my custom flavors and fragrances library gets a hit on the nonsilylated unknowns (I assume it was collected on Chemstation MSD).

[KM-USA]

You can email the GC/MS data to myself sschoenfeld[at]innovaflavors.com and I'd be happy to see if my custom flavors and fragrances library gets a hit on the nonsilylated unknowns (I assume it was collected on Chemstation MSD).

I did send an E-mail. Yes, on Chemstation MSD. I believe you’ll need to make a Directory like UNKNOWN.D like on a flash drive and copy the attached files to that, then load up UNKNOWN.D into your Data Analysis. Thanks for the offer; I think if one gets identified that the series falls into place. I'm OK - if you have colleagues who are intrigued by stuff like this or have other search libraries - if you share the files.

[GOM]

Hi

I find myself drawn back to this one as an interesting academic exercise - or to quote Blackadder to the jailer "The long winter evenings must just fly by"

If we make some assumptions that

1. The underivatised spectrum does not have the background subtracted
2. The disappearance of the unknown peak is due to silylation
3. That ion 104 is the molecular ion

Then apply the following reasoning

1. The molecule contains one of the following functional groups that can be silylated

a. SH - I will discount that on the grounds of unlikely given the context - and because I can
b. NH - I will discount because that would give an odd numbered molecular ion unless there were an even number of amine groups and because the peak does not appear to tail, which is what I would expect on a non-polar column.
c. a COOH, which again I will discount because of a lack tailing on a nonpolar column and lack of ion 60 and the NIST library would have all of those anyway
d. A single trimethyl silyl addition should show a peak in the derivatised sample with a molecular ion of 176 (or 248 if 2 TMS groups are added) - somebody please check my maths.
e. an OH group - hmmm

So assuming an OH (gulp)

I am drawn to an empirical formula of C5H12O2

I have tried various combinations of glycols and glycol ethers but can't get the right fragments or MW

However, thinking about the manufacturing process with ethylene oxide I am drawn towards ethoxylated propanol/ipa CH3-CH2-CH2-O-CH2-CH2-OH. The fragments would be consistent (interesting that TMS would also give a 73 ion but that may be a red herring)

The peak with the 102 ion could be the unsaturated version but I would expect it to elute earlier than the 104 MW on a non polar column?

Anyway it was a fun exercise (the long winter evenings........... ) - please let us know how you get on

Regards

Ralph

[MSCHemist]

I checked it with my libraries and didn't get any hits either. Is it possible the EG is forming an acetal with an aldehyde or ketone. We make some of our flavors in propylene glycol and see that it forms acetals particularly with aldehydes such as vanillin PG acetal or neral geranial pg acetal nonanal etc.

[GOM]

Hmm it is a good thought but EG, being a dihydro alcohol, I think that in a reaction with an aldehyde to form an acetal we would end up with a MW that is too large. Assuming that the MW is 104!!!!

I hadn't considered ethoxylation of an aldehyde

Having retired I don't have access to a NIST library any more so I only have a pen and a scrap pad to work with

The thing that bugs me is that if it is as low as a 104MW molecule it is highly likely that it would get a hit - so we must be looking at something unusual and capable of being derivatised by TMS

What are the impurities in EG? My instinct is to favour ethoxylation of something

Ralph

[rb6banjo]

I'm not sure if they have a restriction on this for outside the US but I use it all the time:

http://webbook.nist.gov/chemistry/

Might be good for the retiree who still likes to dabble in this junk.

The best MW = 104 that I can see with this sort of fragmentation pattern is:

HOCH2CH(CH3)OCH2CH3 (C5H12O2)

It's a monoethylether of propylene glycol. I read in McClafferty's book that primary alcohols tend to give a pretty good MW-18 (in this case 86). This molecule fits the bill in that sense as well.

Unfortunately, the dipropylene glycol analog with the ethyl ether on one end and a primary alcohol group on the other end (C8H18O3) is not in the NIST database that's online. The one that shows the secondary alcohol on one end is there but the fragmentation pattern doesn't match underivatized unknown #2.

[GOM]

Thanks for the link rb

I can get the right fragments with your suggestion as well

So you have monoethyl ether of propylene glycol and I have monopropyl ether of ethylene glycol. The relative abundances don't look quite right in both cases.

hmmm



Cheers

Ralph

[rb6banjo]

Agreed. Could the differences be a function of the mass spectrometer? They are certainly candidates that could be synthesized (or purchased) to see if they match up with the retention time(s) and/or TMS derivatives of the unknowns.

[JMB]

I think that there are some clues here from m/z values that are (A) seen as well as those that (B) are not seen.

(A) IF the MW of underivatized is 104, then m/z 103 is characteristic of loss of H from aldehyde.

(B) My recollection of TMS ethers is that the molecular ion region is usually characterized by BOTH M+. and [M+. - Me.]+, but I cannot see any such pairs. Would expect to see m/z 176 & 161 for a mono-TMS ether of MW 104.

Suggest CH2(OH)CH2CH2OCH=O ????? if MW 104 is true.

At the end of the day, it all rides on knowing the MW's unequivocally.
[EDIT: I would really like to see spectra summed across the chromatographic peak with some background subtraction !!]
Regards,

JMB

[GOM]

Suggest CH2(OH)CH2CH2OCH=O ????? if MW 104 is true.

I can get the right fragments from your suggestion and an aldehyde can be derivatised ( assuming that its disappearance is due to derivatisation) . As I said earlier the MW of 104 is an assumption


Would expect to see m/z 176 & 161 for a mono-TMS ether of MW 104.

You agree with me - phew

At the end of the day, it all rides on knowing the MW's unequivocally.
[EDIT: I would really like to see spectra summed across the chromatographic peak with some background subtraction !!]

Agreed, I made that point earlier

At the end of the day, I have also assumed that this molecule is not present in the NIST database, otherwise it would have been identified by the software

Ralph

[ods-at-pacific]

TMS derivatives of primary aliphatic alcohols show ion current at m/z 73, 75, 89, and 103. TMS derivatives of secondary aliphatic alcohols show ion current at m/z 73 and 75 as well as m/z values due to the substitution on the carbinol carbon. In the the case of these two spectra, this would be the peaks at m/z 103 (89 + 14 = R(CH3)COH) and m/z 117 (103 + 14, ibid).

Both spectra of the TMS derivative exhibit ion current at m/z 147 ((CH3)2Si=OSi(CH3)3). This indicates that more than one group has been derivatized such as a vic di-ol (like whould be the case in the spectrum of a TMS derivative of ethylene glycol or as in the case where to derivatizable groups are located on a chain so that the gas phase molecular ion can fold to the point that the two groups are adjacent to one another as is the case with 1,10-decanediol). Both spectra also show some ion current at m/z 89 (not much, but some); therefore, there could be one methyl substitution and one primary alcohol group.

The non-derivatized spectrum of Unknown 2 (5.302 min) exhibits ion current at m/z 102, which could represent the odd-electron ion [M - H20] (m/z 120).

I think both spectra are secondary alcohols with two OH groups, where one or both have a CH3 on the carbonol carbon(s).

Good luck