Probably you have a sensitive group that's is cleaved during mass spectrometry. For instance, fragmentation sometimes occur on the C-C bond next to the oxygen of an ether bond which results in -15 amu (CH3~CH2OCH3, where ~ is the cleaved bond). Same on alkanes that -CH3 group can be cleaved (e.g. in the mass spectrum of hexane -CH3 can be cleaved and result in -15 amu)

Mr. Joshi,

This which you have received as a final MS corresponds to a summed up MS spectrum. So the deviation of the value would be a result of a fluctuation due to the analyzer.

Open your raw file using this software: https://chemdata.nist.gov/mass-spc/amdis/downloads/ and monitor the value per time. Thus, you carry out a statistical analysis of the datasets allowing you to exclude from a selected contributions to the final m/z value (NB! The procedure should affect the intensity so that there is unable to quantify intensity).

I simulated the 3 mass spectra: C30H27N5O, and two possible losses, mass 16, O or CH4, respectively

C30H27N5 Monoisotopic mass: 457.2266

C29H23N5O Monoisotopic mass: 457.1903

Although the last one is more likely, the difference with the calculated mass is still large. The isotopic pattern is not very differnt ot decide.

The simulated data are attached.

NOTE: "amu" is oblsolete !!!

Mr. Gal,

You can determine unambiguously on the base on the intensity of the cluster of MS peaks experimentaly by means of quantificantion of the experimental intensity, not by means of monitoring of the peak position.

You can pay attention to the following book:

1. https://www.grin.com/document/425061

The method is our development and untill now is tested on 7 classes of compounds. In the manuscript above there is shown 1 class. Others are under consideration currently. It is applicable to CID, ESI, APCI, and MALDI, as well.

You can use also the following approach:

2. Quantitative correlations between collision induced dissociation mass spectrometry coupled with electrospray ionization or atmospheric pressure chemical ionization mass spectrometry – Experiment and theory

Journal of Molecular Structure, Volume 1157, 5 April 2018, Pages 492-512, Bojidarka Ivanova, Michael Spiteller

2. Quantitative collision induced mass spectrometry of substituted piperazines – A correlative analysis between theory and experiment

Journal of Molecular Structure, Volume 1149, 5 December 2017, Pages 243-256, Bojidarka Ivanova, Michael Spiteller

Prof. Bojidarka Ivanova,

I am not sure what you mean by

"You can determine unambiguously on the base on the intensity of the cluster of MS peaks experimentaly by means of quantificantion of the experimental intensity, not by means of monitoring of the peak position."

You CANNOT decide from the isotopic pattern ONLY when there is only one carbons or one oxygen difference in a relatively large molecule or ion. The difference is TOO SMALL! The precise mass (the "peak position" ?) is the first indication of a correct isotopic composition and even the precise mass of the M+1 and M+2 of the ions will be of great value to confirm or infirm the hypotheses of oxygen or methane loss. Good m/z calibration is a prerequisite (but a shift of 16 u is just crazy...). Evidently other techniques like CID, APCI or MALDI will add complementary information. The conditions of ESI may be also critical for observing the quasi-molecular ion (solvent, T, addittives...).

Mr. Gal,

To distinguish between the discussed species by mass spectrometry, quantitatively, does not represent any analytical challenge, currently; furthermore, accounting for the fact that, the assignment is unambiguous, as I have mentioned before.

It is also noteworthy that is not needed a so superior instrumental characteristics, for instance, ultrahigh resolving power at 5-6 decimal signs. With this statement, of course, I do not claim that the resolution does not represent among the most important characteristics of a MS instrumentation. However, several aspectrs of our contributions to development of analytical methods for quantification of experimental MS parameters, show that it can be carried out highly reliable using balanced instrumental characteristics. Simply speaking, there is not needed to use latest generation ICRMS, in order to obtain highly reliable analytical-structural information from the perspective of the quantitative analytical chemistry.

In order to show the validity of the latter statement, a less known theoretical concept should be presented herein (Attachment 1). There are included two of our examples of analyses, which I hope encompass all comments/remarks in your posting illustratively (There are Cl- and Ag-containing analytes and, namely, the isotope effect is taken into consideration). Towards, the remark “complexity” of MS shapes in ESI, due to solvation effects, I hope that via our method, you shall understand that from the point of view of chemometrics a “rich” MS profile represents a significant advantage for quantitative purposes, rather than a limitation. Moreover, our method (on the contrary, the known ones) also operates with MALDI spectra, which is a solid-state approach.

In general, we have developed few methods for quantification of MS intensity. Attachment 1 contains data on our stochastic dynamic approach (Ref. [1]). The approach in Refs. 2 and 3 (my latter posting) is based on Iribarne–Thomson’s theory, which is well-known strategy for quantification of MS intensity. Nevertheless, we have demonstrated, that this strategy can be further exploited for quantitative purposes, studying reaction kinetics. As therein can be seen the correlations with 3D structure are also unambiguous in quantitative terms.

However, let us focus the attention to our stochastic dynamic approach, because of it is applicable not only to different MS methods, but can be used operating to single and/oir tandem modes. On the contrary, the approach in refs. 2 and 3 (above) only operates to tandem mode. Thus, page 1 (Attachment 1) shows our model equation (Eq. 1) and second (known) approach called “current monitoring method” (This model we use in order to verify the validity of our equation (1)). These approaches connect experimental temporal behavior of MS intensity of analyte ions with “diffusion” as an experimental parameter.

It is well known, however, a relation to “diffusion” and “energy” of the molecules, within the Arrhenius’ theory. The energy reflects given 3D molecular and electronic structure of a molecule/ion, which has unique value. Having experimental MS diffusions, therefore, we ca obtain information about 3D molecular and electronic structures computing quantum chemical diffusions. The quantum chemical methods allow us to obtain highly accurate data about any molecule, accounting for solvent, isotopes, ionic strengths, T, P, and more, factors/parameters.

In the first part of Attachment 1 there are provided experimental MS spectra, parameters and calculations of diffusions according to equation (1). In addition, there are shown data according to current monitoring method (Equation (2a)). As can be seen there is obtained coefficient of correlation r > 0.99. We have examples when r > 0.999, studying 18 characteristic MS ions per analye within a relatively small range m/z-values. Therefore, the correlations are statistically meaningful.

The next part of Attachment 1 focuses on correlations with experimental MS and quantum chemical diffusions, obtained according to the Arrhenius’ approximation. As can be seen r is equal to 0.96049. Nevertheless, that we have obtained better statistical parameters studying a large number of other analytes (submitted for consideration), even this value of (1) is enough representative to a real linear correlation with experimental stochastic diffusions according to our model equation (1) and the Arrhenius’ formalism (equation (5)).

Once when, for a given analyte, there are obtained statistical equations connecting experimental diffusions on the base on our stochastic dynamic approach and (eventually!) theoretical quantum chemical diffusions (Despite the fact, that there can be employed only equation (1), using a set of statistical relations of reference samples), any analyte is determined unambiguously even in complex mixture on the base on: monitoring of intensity of isotope cluster of peaks. Macromolecules can be also determined using the characteristic fragment MS ions at m/z < 900, because of, any molecule has its unique fragment pattern at low m/z values. Here, of course, the superior resolving power cannot be neglected, because of, there is expected a decreasing in chemometric performances in cases of MS spectra of proteins measured of low resolution.

As far as the quantum chemistry, via computation of vibrational frequencies provides very sensible information about energetics, you can distinguish not only between the discussed molecular fragments, but even between isotopes of one and the same fragment. In this context, attachment 2 contains frequencies of 107AgCH3CN+ and 109AgCH3CN+ complexes. As can be seen, the data are remarkably different in quantitative terms. Therefore, the theoretical diffusions, like the corresponding experimental parameters, are also very distinguishable quantitatively.

You can test the universal validity of our equation (1), using the procedures shown as attachment 1, citing ref. 1, of course. The software for processing of MS spectra is shown in my latter posting. It is free of charge and operates with original raws, among other file-formats. The curve fitting using SineSqr approximation can be carried out employing any statistical software. The second part of the analysis can be done using, for example, GAMESS-US (http://www.msg.ameslab.gov/gamess/) (It is free of charge). If you have Gaussian (www.gaussian.com), the analysis should be faster, because of, the larget part of the computational steps are almost completely authomatized.

I don't see how your model may resolve the question (loss of 16).

I just state that there is some uncertainties in intensities due to noise, instrumental factors, etc. The precision on m/z even with 4 decimals is already good for eliminate a number of hypotheses. I never mention ultra high resolution and precision like in FTICR.

Mr. Gal,

One of the greatest advantages of our quantitative methods, amongst others, consists of the fact that we quantify absolute intensity.

The attachment 3 shows typical order of magnitudes of absolute intensities and noises of ESI and MALDI spectra.

For example:

Intensity of a ESI peak: 2 869 854.5 (Noise 676.09)

Intensity of a MALDI peak: 75 223 848.0 (Noise 1).

The shown noise per MS peak is a total-average value, which varies over the scan-time range, of course, and thus these (frequently negligible) “uncertainties” should be taken over all ranges of measurement times.

For example, looking at the MALDI data in ref. 1, the noise varies from 1 to 163 in the whole range of times. Let us asses the uncertainty, using the standard deviation, nevertheless that it does not reflect exactly the measurement uncertainty using ISO 5725-1, but is very illustrative about the capability of our method:

Mean: 7.52239E7, SD(yEr+/-): 29.34138 and SE(yEr+/-): 14.67069, respectively.

Therefore to the mean value 75 million abs units, there is an average contribution due to noise 29 abs units. Such contribution to the MS intensity affects insignificantly on performances, in context chemometric tests.

An analysis, based on I = 75223848 +/- 29 units represents, as I have mentioned before, an absolute determination and a unambiguous assignment of an analyte.

In addition, I should mention the fact that the noise, studying isotope shape contributes equally to the absolute intensity. The variation between the noise values to each of the component of peaks of the isotope shape is at 1-2 decimal sign (Attachment 4). This fact leads to an unambiguous determination of different isotope compositions, like for example, mono- and/or dinuclear complex species, and more.

On the other hand, looking at our equation (1), the error contribution is determined on the base on the constant lnP1 = -17.053(4). As you can see from the data in the already shown tables, this determines an accuracy at the 4 decimal sign. Therefore, there is a lack of error contribution to the so–called MS diffusion parameter based on the model equation (1).

My question is "can you resolve the question related to loss of 16". Can you use the data given by Gaurav Joshi or my hypothetical C30H27N5 & C29H23N5O to resolve the case ?

Mr. Gal,

The answer to your question is yes! Because of as you can see we have developed universal method, meaning - law. Moreover it is tested on over 150 compounds.

In this context, upload a MS spectrum of any compound. Just, please not very large molecule, because of, if it is large, then it should take more time for computation at high accuracy. For example, if you have a spectrum of a LWM pharmaceutics...it would be good (For example, paracetamol or similar compound, meaning at 100-150 atoms). There is needed original raw. The fragment pattern sould contain the topic of your question (or doubt towards fragment pattern). In addition, let the MS spectrum be of organics, because of the computation of inorganics also takes more time. As well as, let the measurement exclusdes from internal standard, thus, it shall be lack of doubt towards the capability of our method.

In general, I would provide you original data, however our objects are very complex - peptides, dyes, metal-organics, etc. We have not so simple and illustrative example, which to include the topic of your question.

Again it is not my question, it is the question of Gaurav Joshi. I tried to answer his problem. I don't see a conrete answer to his question.

Mr. Gal,

The concrete answer is yes as I have writte before!.

It is shown still in my first attachment. As I have written several times as well as the quantitative computation of the diffusion parameters, which reflects exact structures and its correlation with MS diffuusions, according to our law (equation (1)) provides unambiguous determination. Why? Because of, if you have, for example, MS spectrum p-OH-aniline and has doubt towards loss of OH or NH2, in the former case the fragment product is aniline cation, while in the latter case the product MS ion should be phenol-cation. The energetics of the phenol and anilite cations is significantly distinbuishable quantitatively.

The quantum chemitry and it laws are known from more than 100 years. The energetics and thermodynamics of chemical reactions are basic knowledge from BSc-grad in the Chemistry.

With affiliation 'analytical chemistry' and higher education in chemistry you are familiar with the laws of the quantum chemistry and thermodynamics.

Given that, would you like to test independently the validity of our new quantitative law providing a raw or just take part in a discussion .....?

In addition, if the discussion is initiated by Mr. Joshi, why you engaged so intensively with it?

You did not answer to my proposal: is the loss O-16 or CH4?

I engaged the discussion with you because you answered directly to me.

I beleive that your model may be useful in general and that the theory is sound. But if the structure of C30H27N5O is not known, how to apply it?

Concerning the basic knowledge of chemistry: I am 73, emeritus and I taught MS for about 20 years. I taught also thermochemistry, spectroscopy, physical organic chemistry...

Mr. Gal,

The capability of our methods is more than highlighted in context of your latter posting. But, as I have mentioned, we have not so simple molecules. Neverheless, I am providing you a remarkable case, which to illustrate the capability of our quantification approach. It shall be even more interesting because of, is an example of nontrivial MS behavior of a peptide. It is object of our projects and is the most simple peptide in our dataset of spectra.

In order to spend time in writing, I am uploading you a text and supporting information file to our (also) submitted for publication work. It is uploaded the original raw, as well. So that you can follow the data in the tables and figures using the original dataset.

Direct answer to he question: ''How to distinguish between similar molecules is quantitatively using mass spectrometry?'' is: using the frequency analysis at a stable and a transition state computing quantum chemical diffusions according to eq. (13) in text. Any structural change affects significantly the frequencies and thus the theoretical D parameter. Such structure is excluided from the linear relationship: experimental diffusion = f (theoretical diffusion) (Fig. 7 in the text).

In the latter context, I am uploading two files of MS ion at m/z 271. The frequencies in GS are all positivem while in TS there is one negative frequency.The latter is neglected. At this point we follow the computational approach in ref. [18], because of there is debate in the literature towards the exact expression of equation (13), but our experimental data show that to the experiment fits to the description in:

[18] A. Nitzan, Chemical Dynamics in Condensed Phases, Oxford University Press, Oxford, 2007, pp. 1–719

This paper, I hope explain maximum well the computational/calculation steps and the capability of our new quantitative law in the mass spectrometry.

Meanwhile, I shall compute theoretical diffusions of hypothetical ions at ramge of m/z values shown in Fig. 7 (structural isomers, for example) and at about a week I shall upload these data, as well. Thus, it shall be demonstrated how can be distinguished between, even, isomers, not different structures, quantitatively, using our method. There is needed only conventional MS and quantum chemistry.

I don't doubt that your apporach is useful in general for the interpretation of mass spectra. It was not my point in the previous answers.

I'll try to read your documents but it will take a while before I could assimilate your model. In my opinion, it would be more productive to ask diretcly Gaurav Joshi to give more information on his molecule. I fear that he was not really interested in our conversation...

Dear Prof. Jean and Ivanova, I am quite active in going through your conversations and getting the worthy intake on my topic. Actually this is matter of high privilege that I am getting to know more about the technicalities, vast research inputs and more importantly the fundamentals on the theme, you discussion is apparently helping me a lot to grasp the subject in more details.

Considering the more information on my molecule there is a methyl group directly attached to carbonyl group which I suspect that might cleave to give a loss of 15, but actually there coming a loss of 16 which is worrying me. Whether what could be the possible explanation for this.

I will also be uploading the raw file of my HRMS file once I will get same from the technician.

Anyway this was the healthy discussion I just learn lot many stuff.

Much Appreciable!

Thank you

Mr. Joshi,

Looking at your attachment, once when you obtain your raw follow the calculations and computations in my latter posting. If you have not available Gaussian (www.gaussian.com) contact someone in your University familiar with the quantum chemistry. The molecule is not so large and there should not be problem to obtain the data fast. The equations are shown in the submitted paper together with the reference section, so that the Arrhenius' fomalism within the framework of TST is well known and you can find it well explained in a lot of books and textbooks.

The attachment are deleted due to accaptance of the work. The rewritten version of teh work should be available via the regular providers of the scientific literature.

Concerning your remark in your last post:

"Considering the more information on my molecule there is a methyl group directly attached to carbonyl group which I suspect that might cleave to give a loss of 15, but actually there coming a loss of 16 which is worrying me. Whether what could be the possible explanation for this."

Don't forget that you are dealing - in principle - with a protonated molecule in ESI+. Loss of a methyl radical to get a R-C=O+ is valid for electron ionization. For a protonated ketone, loss of methane gives also the stable acylium ion. From the spectrum and precise mass, it the most probable interpretation.

On the other hand, the absence of protonated parent molecule in the spectrum is puzzling. You should try (i) changing the source conditions to get less collisional activation (ii) try another ESI solvent, (iii) try to get a Li+ or Na+ adduct by adding a salt to the solution: if the molecule has some affinity for the cation, you'll see the adduct M+7 or M+23.

Thank you for the kind advice Prof. Gal. I have referred to your point (i) and got my desired mass.

Good ! Note also that the isotopic pattern (see the mass spectra simulations) is useful confirmation, in addition to the precise mass.