An alternative look at ink wetness

[InesF]

A short update about ink wetness measurements.

 

At first a new method

Fountain pen ink density is measured by weighting a defined volume of ink. A 10 mL volumetric flask was used, containing 10.00 mL volume (calibrated with deionized water). Empty and filled weights were measured with a Carbon Express electronic balance, Item No. CS-59-II, which has a range of 100 g and a precision of +/- 0.02 g. Filling 10 mL ink results in a precision of +/- 0.002 g/mL. After each measurement, the volumetric flask was cleaned (rinsed with deionized water until colourless) and rinsed two times with 1 mL ink (to remove residual water).

 

The data set

Meanwhile 42 fountain pen inks are measured for their properties and for ink delivery/consumption on two paper types (absorbent and 'ink friendly') with two different fountain pens each. 23 more inks are measured for their basic physical and chemical properties. As the measurement of density started recently, the related data set is currently small, but will be extended to all 42 inks in due time.

 

The updated data table is not much different from last update, only one ink (Lamy Crystal Beryl) is added. Visual impression of the ink amount laid down on a paper is based on volume and not on weight, therefore ink density measurement is now added. At this stage, ink density doesn't seem to vary a lot. However, a volume based principal component analysis may be a bit different from the (current) weight based.

 

The new principal component analysis with 42 inks did not show much differences to the previous with 32 inks. Factor loadings (arrow lengths) of conductivity and viscosity have both changed a bit, becoming similar lengths now. The ink line widths are now, on basis of 42 inks, grouping a bit better while all other factors remain almost unchanged.

 

Therefore I can repeat and refine the conclusions

- Higher surface tension reduces the amount (weight) of ink laid down to the paper

- Higher viscosity reduces the line width and slightly increases the amount of ink laid down

- Higher viscosity does NOT reduce ink flow of a fully functional fountain pen

- Higher viscosity sharpens the edges of the ink line and let it appear more crisp and more saturated

- Both, higher viscosity and higher conductivity do reduce the ink line width a little bit

- Lower surface tension supports spreading and feathering of inks on absorbent paper

- Inks with either lower surface tension or higher viscosity or much higher conductivity visually appear to be wetter

- Surface tension is the most influencing factor on ink wetness, followed by viscosity and conductivity with variance contributions of 63, 11 and 7% respectively

 

Next updates

When density measurements are finished, I will calculate a new principal component analysis for ink wetness which will be volume based (instead of weight based). In parallel I'm contacting some scientific journals, asking for their interest in publishing this new alternative ink wetness theory. Because of that, I will not publish the volume based ink wetness data here at FPN, avoiding conflict of interest. The publication will be open access - no secrets, no paywall, no worry!

 

There is no time schedule for the next update, however, as always, I invite you to discuss and comment.

Thank you!

[LizEF]

Thanks for the update, @InesF!  Good stuff.

[txomsy]

If I may...

 

You may have a flaw in your density measures. The problem comes in the precision of the measure and stems from the measure of both volumes and weights. Without more detail on how it is done, I cannot be categorical, but...

 

You only have one measure. I assume that given the price of ink, it may seem as expensive to afford. Not necessarily.

 

Regarding ink: I don't exactly know how do you measure the amount of ink, so can hardly comment, but if it is by filling in a test tube or similar, the precision would be related to the error in your ability to consistently deploy ink up to a mark, and the precision of that mark. You'd want a hair-thick mark and to match it as precisely as possible. And that is assuming temperature is controlled in all experiments, for it could also affect (albeit slightly) the volume of a fluid. I am assuming constant pressure, but climatology may also play nasty tricks, just consider that it is easy for atmospheric pressure to range between, say 995 and 1024 millibars, and could potentially, in the extremes, range from 885 hPa to 1077 hPa. But assuming one does the measures in the same place and day or controls pressure with a barometer, and at constant temperature, that should not be an issue.

 

To increase precision, you would want the tube to be as thin as possible so the ratio of error in precision matching the mark to the total length is as large as possible (hence reducing the error). But then, you hit a second issue: since surface tension is different, you might end up measuring different quantities just because, due to surface tension, you reach the mark at different quantities. And inks are known to have different quantities of surfactants. Again, if you can find a thin-enough tube made of a material where ink surface tension is minimal, you could increase precision. Can't be sure, but maybe a 1mm syringe would work, as they are thin and made to allow use with different fluids; but the marks are relatively thick; a 0.01ml error in a 1ml measure is arguably not important in Medicine, but might be here.

 

In any case, I doubt most peer reviewers would accept only one measure.

 

You are filling the recipient and washing with ink. I would venture that it is overkill. What you would want is to wash it thoroughly with water and then let it dry completely. So, fill it, pour the ink in, pour the ink out back into the inkwell with the rest of the ink (at this point you would lose some ink attached to the walls), wash thoroughly with water and let it dry completely before filling it. That way there is no contamination and no remnants. And then you could re-use the former load of ink that you put back in the ink bottle with the other ink so that you start a-fresh. This would minimize the use of ink to about 10ml ± 0.1ml/measure and you'd recover the 10ml at the end.

 

Consider that a 0.01 error in the two measures results in a precision of 0.99 either, which combined gives 0.98, i.e. an error of ±2%, which would make your single measures of, e.g. 0.996, 1.000 and 1.006 non-significantly different and the 1.022 might by by chance. Unless I am mistaken (and it would not be rare for me, specially before breakfast).

 

Combining a careful measuring method with a careful reuse of the vase, should allow you to make many measures, and, in this case I would try as many as possible. There was this (half-jest) practical recipe for sample size:

 

    sample_size = amount-of-$$ / cost-of.experiment

 

Next the weighting. If your balance has a precision of 100 ± 0.02, then when you are measuring 10g, your error is still ±0.02. The error is not a proportion but an absolute error in the measure for the range of the instrument. So your estimation seems to me (in principle and in the absence of more detail on the materials and methods employed) overly optimistic.

 

So, I'd say more measures would be needed prior to publication. But I may be wrong as this is not my field.

 

Of course, nothing of this holds if you can use a high precision system for both volume and weight measures, but if I were a reviewer I'd tend to lean on the skeptic side.

 

[InesF]

1 hour ago, txomsy said:

If I may...

You are cordially welcome, @txomsy!

 

1 hour ago, txomsy said:

You may have a flaw in your density measures. The problem comes in the precision of the measure and stems from the measure of both volumes and weights. Without more detail on how it is done, I cannot be categorical, but...

It is a 10 mL volumetric flask that I checked previously with deionised water with the analytical balance. It has a volume of 10.00 mL. The +/- 0.02 comes from the limited precision of my small gemstone balance (the Carbon Express thing).

To be honest, the problem is less the precision of the flask it is much more in filling it correctly. The meniscus should exactly touch the mark, easy with water or liquids of pale colour, but comparably hard with inks.

I'm not sure if I stay with this pycnometer-type of measurement - but currently I have no other option.

 

1 hour ago, txomsy said:

Consider that a 0.01 error in the two measures results in a precision of 0.99 either, which combined gives 0.98, i.e. an error of ±2%, which would make your single measures of, e.g. 0.996, 1.000 and 1.006 non-significantly different and the 1.022 might by by chance. Unless I am mistaken (and it would not be rare for me, specially before breakfast).

My balance precision is +/- 0.02 g - that's why I use a 10 mL flask to bring it down to 0.2%. I'm afraid, this is still not good enough.

... always start serious work after the coffee break! 😄 

 

1 hour ago, txomsy said:

Of course, nothing of this holds if you can use a high precision system for both volume and weight measures, but if I were a reviewer I'd tend to lean on the skeptic side.

Indeed, reviewers are natural enemies of authors ... 

 

Thank you for your comments!

 

[InesF]

10 hours ago, LizEF said:

Thanks for the update, @InesF!  Good stuff.

Thank you @LizEF - your comments always kept my motivation high!

[txomsy]

Thanks for the details, I had indeed misunderstood the original post, which just comes to show I am more dependent on coffee than I'd like to admit.

 

[LizEF]

5 hours ago, InesF said:

Thank you @LizEF - your comments always kept my motivation high!

Glad to have helped.  It can be hard to keep up motivation on such a long project¹, but this is so very interesting that it's easy to "egg you on"!

 

¹There are days when I really don't feel like recording a review (the recording and editing are the hardest/most time-consuming part), but I love the interaction I get - that's what motivates me.  (That and a slightly obsessive need to "finish" sets - and the set of inks on the planet is rather large and I'm nowhere near done...)

[Sholom]

I will add my continued encouragement. This is fascinating and a much appreciated lot of work. Good luck on publication!I will add my continued encouragement. This is fascinating and a much appreciated lot of work. Good luck on publication!

[InesF]

Thank you @Sholom, you're cordially welcome!

[squirrels]

This is a fascinating and valuable project! Thanks for doing it!

[InesF]

Thank you @squirrels! And a warm welcome to the community!

[InesF]

One more update from the ink measurement project.

 

As mentioned before, the initial procedure for measuring density of inks didn't work well. Precision was not good enough.

I changed the method to a pipetting of 1 mL ink (Gilson pipetman, #17894C) into a small baker at the analytical balance (0.1 mg precision) for 5 or 6 times and reporting each mL weight separately. For calibration I did the same 6 to 9 times with water. It turned out, that the absolute volume of the pipette is about 1.5 or 2% below the adjusted 1.000 mL but is reproducible with less than 0.2, often less than 0.1% deviation. That's good enough for density measurements.

 

Here is a 'final' update of ink physical and chemical data:

*) ink without glimmer particles.

 

Fountain pen ink densities do not vary a lot, they are all in the range between 1.002 and 1.046 (currently only those inks are measured which are also measured for delivery/consumption). I will not use ink density as a stand alone variable in principle component analysis but for conversion of ink delivery from grams per line length into mL per line length. The values do not change a lot, but enough to improve the correlation a tiny bit.

 

The following is a raw data table from one measurement day of my ink density measurements. It has data for 8 inks and four rows of water calibration (at the beginning and then after 3 inks).

Mean values (MW), standard deviations (stab), variance in percent (v%) and calculated density are listed in the bottom 4 lines of the table. The pipetted volume was repeatedly checked with water (violet data rows) and density calculation of inks was related to the nearest two water data rows.

 

I forgot to upload the overview photo of ink electrical conductivity measurement. Here it is:

 

There you can see the 200 µL Gilson pipetman (the smaller sister of the one used for density measurement). This small handheld conductivity meter proved to be surprisingly precise and robust.

 

I have no photos from ink density measurement. I used the same analytical balance as shown for ink delivery measurements (second entry, second image of this thread).

 

 

Due to personal reasons, I have to take a break over Christmas.

Nevertheless, I invite you to share your thoughts and discuss the ink wetnes theory. I will be back and reply in January 2022.

 

Have a good Christmas time, all the best for the New Year

and stay healthy!

[LizEF]

Merry Christmas and Happy New Year, @InesF!  Thank you so much for continuing this project - and the fabulous home-scientist photos.

[mtcn77]

Hi @InesF!

Can I ask for a few pointers on the list? Diamine Green Black is pretty viscous on the list. How should we interpret it, it is listed wet on an ink guy review, I must be doing something wrong...

[InesF]

Hi @mtcn77, you're welcome!

 

No, not totally wrong. Diamine Green Black is

1) the most viscous ink out of the bottle I have tested so far and

2) with 64,2 mN/m at the lower end of dry, close to medium wet.

Please be aware that the classification I did in the table has subjective limit values. No doubt, comparing the delivery amount of two inks written with one pen on one paper, a bigger amount of an ink with lower surface tension is delivered. The behaviour difference between an ink with 62,6 mM/m surface tension and another with 64,2 is marginal.

 

According to the statistics, the D. Green Black appears to be wetter (on paper) with less amount of ink delivered to the paper. I didn't test dry time on paper - in case it is long (I guess so) this also contributes to the subjective impression of increased 'wetness'.

 

Nevertheless, D. Green Black behaved comparably dry in my writing tests with two M and one EF nib.

 

Hope I could answer your question. If not, please let me know!

 

[mtcn77]

@InesF wow, thanks!

I'm looking out for it in the driest of my pens, noted. I went on a shopping spree and have some findings I want to report. I did some testing in which wetness in the pen and on the page turned out to be different.

 

According to an ink guy, Parker Quink Black is wetter than Pelikan 4001 Brilliant Black; however on the page, the reverse order is visible: Parker Quink Black is narrower while Pelikan 4001 Brilliant Black is deeper and more feathering. I think it has something to do with ink viscosity.

 

I had success using the Pelikan 4001 Brilliant Black ink in the more moderately flowing pen and Parker Quink Black in the gusher. Both turned out okay. Pelikan(high viscosity) made more moderately flowing pen really smooth and Parker(low viscosity) controlled the slight nib misaligned, hardstarting and skipping of my wet writing pen.

 

Also, they match their swatch better when paired up correctly. Parker was fading and Pelikan was almost sheening in the other pen.

 

Thanks for the scientific conduct, I couldn't find my inks alone.

[arcfide]

When comparing this work on wetness to that of An Ink Guy, remember that An Ink Guy is just measuring viscosity, which is not the same thing as surface tension in these inks. His "wetness" test is related to flow relative to water. If you put plain water in your pen and write with it, you'll get something that probably "feels" quite dry. That's because An Ink Guy isn't defining ink wetness on the basis of how smooth the ink is to write with or how much ink it puts down with a given nib, but just on flow characteristics relative to water. Here ink wetness is defined differently. It's important when discussing ink wetness to determine what you mean by "wet". 

[mtcn77]

7 minutes ago, arcfide said:

When comparing this work on wetness to that of An Ink Guy, remember that An Ink Guy is just measuring viscosity, which is not the same thing as surface tension in these inks. His "wetness" test is related to flow relative to water. If you put plain water in your pen and write with it, you'll get something that probably "feels" quite dry. That's because An Ink Guy isn't defining ink wetness on the basis of how smooth the ink is to write with or how much ink it puts down with a given nib, but just on flow characteristics relative to water. Here ink wetness is defined differently. It's important when discussing ink wetness to determine what you mean by "wet". 

Yes, it should be elucidated his viscosity scale sometimes matches the surface tension scale, in other times the viscosity scale. Where other results are present like @InesF findings, I think surface tension scale is the better characteristic. For instance, Waterman Mysterious Blue looks wet on his results, but dry here. I haven't tested it, but I think it should be dry indeed, if it writes wet.

 

In my humble experiment, dry pen + dry ink = wet pen, wet pen + dry ink = feathering, wet pen + wet ink = dry pen with crisp lines and dry pen + wet ink = dry pen.

I just think what people say as wet and what the tests say as wet are two different things. We might be pointed in the wrong direction, if there is a misnomer in regard to surface tension which when high should make the ink spread by making a big ink bubble at the tip and should not fully lubricate the tip when low, since low tension ink should not bead up the same amount as high surface tension.

In summary, there is a misnomer, in my opinion.

[InesF]

1 hour ago, mtcn77 said:

In my humble experiment, dry pen + dry ink = wet pen, wet pen + dry ink = feathering, wet pen + wet ink = dry pen with crisp lines and dry pen + wet ink = dry pen.

Hi @mtcn77, I have to admit, that reads a bit like a mathematical double integral of a summary function to me. 👩‍🎓 But your observation is an interesting surprise!

 

The thread has the word 'alternative' in it (would had been better titled 'alternative theory', but never mind). In contrast to the subjective impression of wet and dry - which I can't measure - the amount of ink laid down by a fountain pen is something much more objective. And that's exactly what I measured with my experiments. In this context, I can differentiate between wet (much ink per line surface) and dry (less ink per line surface) behaviour. The evaluation of measurement results led to the summary: an ink is wetter if it has lower surface tension and if it has higher viscosity.

While the first part is obvious and in line with our sensual impression, the second part is counter-intuitive and is opposite to the 'common' theory. I tried to find an explanation in the description of the drawing (some pages above).

 

Since I adapted to those new findings, it never happened again to me to chose an 'alternatively' dry ink for a dry writing pen (no more pale looking lines) and to chose an 'alternatively' wet ink for a wet writing pen (no more ink floods on paper with all its consequences).

Two examples: deAtramentis Virginia Woolf ink (39 mN/m) works great with a Montblanc EF and causes ink floods with any F, M or B Pelikan nib, while Waterman Audacious Red (71 mN/m) has tamed my wet gusher TWSBI Vac700r to draw a perfectly medium-wet but colourful, rich line.

 

What makes a pen wet or dry is a secret of the manufacturer. I have a hypothesis - proofing it will keep me busy in 2022. 😅

[mtcn77]

2 hours ago, InesF said:

Hi @mtcn77, I have to admit, that reads a bit like a mathematical double integral of a summary function to me. 👩‍🎓 But your observation is an interesting surprise!

Well, hello there!

 

(You're supposed to quote as, - General Kenobi!)

 

Yes, it is. i have a theory why that seems so counter-intuitive, though: it is because we are accustomed to a simple answer. Allow me to elaborate;

 

Since test results and subjective pen smoothness feeling have different findings we are accustomed to resorting to our own judgement, mixing them up in the process. For instance,

 

2 hours ago, InesF said:

The evaluation of measurement results led to the summary: an ink is wetter if it has lower surface tension and if it has higher viscosity.

Technically, these should be proportional to wetness, but subjectively inverse to a wet writing pen. I couldn't comprehend one part in this though, in what context do you use;

On 11/28/2021 at 1:23 PM, InesF said:

- Lower surface tension supports spreading and feathering of inks on absorbent paper

Does it mean to increase feathering, or to decrease it?

 

I'll fill in some more quotes;

On 11/28/2021 at 1:23 PM, InesF said:

Therefore I can repeat and refine the conclusions

- Higher surface tension reduces the amount (weight) of ink laid down to the paper

- Higher viscosity reduces the line width and slightly increases the amount of ink laid down

- Higher viscosity does NOT reduce ink flow of a fully functional fountain pen

- Higher viscosity sharpens the edges of the ink line and let it appear more crisp and more saturated

- Both, higher viscosity and higher conductivity do reduce the ink line width a little bit

- Lower surface tension supports spreading and feathering of inks on absorbent paper

- Inks with either lower surface tension or higher viscosity or much higher conductivity visually appear to be wetter

- Surface tension is the most influencing factor on ink wetness, followed by viscosity and conductivity with variance contributions of 63, 11 and 7% respectively

Here, you are noting higher viscosity decreases line width and higher surface tension decreases ink deposition, but you never say higher surface tension 'increases' line width - this my finding so far.

 

I also agree lower surface tension controls feathering(if that is what you mean), what is not stated is that higher surface tension 'increases' feathering  - also what I found on cheap paper. As these two statements contradict the "high surface tension = dry ink" preposition, I think there are still more ink properties left to uncover.

 

Where the complication stems from is due to the interchangeable context of 'high viscosity' inks with 'high surface tension' inks - those aren't the same property, but mixed up with one another. Like you said it,

On 11/28/2021 at 1:23 PM, InesF said:

- Higher viscosity reduces the line width and slightly increases the amount of ink laid down

Lines will appear narrower since high viscosity('saturation' as I conceptualise it) makes the ink 'drier'. High surface tension on the other hand - which might be mistaken for "high viscosity" - will instead make the pen write wider, in my opinion.

 

We can test it out, I'd just ask you to write the same summary for high surface tension inks to differentiate them from your previously stated high viscosity inks.