Hansen Solubility Parameters and Their Relevance to Dyne Testing

Question:

I’ve been told that some polymers cannot be evaluated with the dyne test, and I’m especially concerned about any solubility of the plastic in the test ink. Can you give any guidelines on when the ink may interact with the surface, and explain how to tell if this is a problem?

Answer:

A simple rule of thumb is that if a solvent causes any melting of the polymer surface, or swelling of the polymer bulk, there is a degree of solubility present. Most published data on solubility look at how resistant the solid is to the solvent over a long period of time – this is important for prediction of durability and suitability for most commercial applications. But whether or not the material interacts over a period of hours or days or if it interacts only at elevated temperatures is not really relevant to the dyne test, which is based on the solid—liquid interaction over a time frame of only a couple seconds, hopefully in a standard laboratory environment.

The specific answer to your question is that it really comes down to discretion: Is there evidence of solubility within several seconds, or does it take minutes or longer to see any swelling or melting of the surface? My personal feeling is that if the surface remains unaltered for at least about 10 – 15 seconds (a scuff test or similar would demonstrate this empirically) you are probably OK, though I would consider the results to be somewhat tentative. In other words, if there is some evidence of solubility within, say a quarter to a half minute, and there is any other reason to believe the dyne test results may be inaccurate, that would be a concern. If the solubility is apparent more quickly, I would not trust the results at all.

You question has piqued my curiosity, so I’ve done a more scientific look at the issue, using Hansen Solubility Parameters (HSP) as the basis for the investigation. In HSP terms, the affinity between two molecules is considered as a 3 dimensional distance, defined as(1)

Ra2 = 4(δD1 – δD2)2 + (δP1 – δP2)2 + (δH1 – δH2)2    (1)

where in this case δD1, δP1, and δH1 are the dispersion, polar, and hydrogen HSP components, which define the location of the molecule in “Hansen space” for the polymer and δD2, δP2, and δH2 are the HSP components for the solvent. The Relative Energy Difference (RED) is defined as Ra/Ro, where Ro is the radius of the sphere of influence (interaction radius) for the polymer. Any solvent which lies less than the RO away from the centerpoint of the polymer’s sphere will allow solubility of the polymer in that solvent; thus, an RED of less than 1.0 suggests that some degree of solubility will occur. The Ro of different polymers varies quite a lot; the smaller it is, the more precisely the solvent’s solubility parameters must match those of the polymer for solubility to occur. A higher molecular weight version of a given polymer will have a smaller sphere of influence, and be less readily soluted; the same holds true for a crystalline, as opposed to amorphous, molecular structure(2).

Based on this conceptually simple (though not so easy to wrap one’s head around) model, it is a straightforward process to calculate whether or not a polymer will be soluble in a solvent, including those used in the formulation of dyne solutions. Table I shows the Hansen parameters for the three constituents of the test fluids(1). Table II shows the Hansen parameters for the formulations of all the even levels from 30 through 72 dynes/cm, as determined by their volumetric percentages. Table III shows the critical surface energy and Hansen parameters for a variety of polymers.

TABLE I
Solvent Dispersion Polar Hydrogen
2-ethoxyethanol 16.2 9.2 14.3
formamide 17.2 26.2 19.0
water 15.6 16.0 42.3
TABLE II(a)
Dyne Level % 2-ethoxyethanol % formamide %water Dispersion Polar Hydrogen
30 100.0 0 0 16.2 9.2 14.3
32 89.5 10.5 0 16.3 11.0 14.8
34 73.5 26.5 0 16.5 13.7 15.5
36 57.5 42.5 0 16.6 16.4 16.3
38 46.0 54.0 0 16.7 18.4 16.8
40 36.5 63.5 0 16.8 20.0 17.3
42 28.5 71.5 0 16.9 21.4 17.7
44 22.0 78.0 0 17.0 22.5 18.0
46 17.2 82.8 0 17.0 23.3 18.2
48 13.0 87.0 0 17.1 24.0 18.4
50 9.3 90.7 0 17.1 24.6 18.6
52 6.3 93.7 0 17.1 25.1 18.7
54 3.5 96.5 0 17.2 25.6 18.8
56 1.0 99.0 0 17.2 26.0 19.0
58 0 81.2 18.8 16.9 24.3 23.4
60 0 65.0 35.0 16.6 22.6 27.2
62 0 47.0 53.0 16.4 20.8 31.3
64 0 30.6 69.4 16.1 19.1 35.2
66 0 18.2 81.8 15.9 17.9 38.1
68 0 8.6 91.4 15.7 16.9 40.3
70 0 3.6 96.4 15.7 16.4 41.5
72 0 0 100.0 15.6 16.0 42.3

(a) Formulation data for dyne levels 30 through 56 is from ASTM Std. D2578(3). Values for dyne levels 58 through 72 are not specified by ASTM – the values are from actual ACCU DYNE TEST formulations.

Table III(a)
Polymer Name CST(b) Dispersion Parameter(c) Polar Parameter(c) Hydrogen  Parameter(c) Ro(d)
Acrylonitrile butadiene styrene 38.5 17.0 5.7 6.8 9.4
Cellophane 45.4(e) 16.1 18.5 14.5 7.3
Cellulose acetate 37.5 17.1 13.1 9.4 10.6
Cellulose acetate butyrate 34.0(e) 17.2 13.8 2.8 12.6
Epoxies 44.5 19.2 10.9 9.6 11.1
Ethyl cellulose 30.3 19.0 5.6 4.9 7.9
Fluorinated ethylene polypropylene 19.1 19.0 4.0 3.0 4.0
Nitrocellulose 42.7(e) 16.2 14.1 9.5 10.7
Nylon 6 43.9 17.0 3.4 10.6 5.1
Nylon 6,6 42.2 17.2 9.9 16.5 4.4
Nylon 11 35.6 17.0 3.4 10.6 5.1
Nylon 12 37.1 18.5 8.1 9.1 6.3
Poly a-methyl styrene(f) 38.7(e) 18.5 2.4 2.4 (g)
Poly iso-butyl methacrylate 30.9(e) 18.3 4.9 6.3 9.4
Poly n-butyl methacrylate 29.8 16.0 6.2 6.6 9.5
Polyacrylonitrile 47.0 21.7 14.1 9.1 10.9
Polybutadiene 29.3 17.5 2.3 3.4 6.6
Polycarbonate 44.0 19.6 8.8 5.7 10.2
Polychlorotrifluoroethylene 30.8 15.6 2.5 4.7 5.8
Polyether sulfone 47.0(e) 18.3 8.2 6.4 6.3
Polyetherimide (g) 17.7 6.0 6.4 4.8
Polyethyl methacrylate 32.8 17.5 8.5 4.7 9.6
Polyethylene 31.6 16.8 3.8 3.8 6.6
Polyethylene oxide(h) 43.0 17.0 10.0 5.0 (g)
Polyethylene terephthalate 39.0 18.7 6.3 6.7 6.5
Polyisobutylene 27.0 16.4 1.7 4.7 7.9
Polyisoprene 32.3(e) 18.1 2.4 2.3 10.3
Polylactic acid(i) 38.0(j) 18.5 8.0 7.0 (g)
Polymethacrylonitrile 39.0 17.2 14.4 7.6 3.8
Polymethyl methacrylate 37.5 17.9 10.1 5.4 11.0
Polyoxymethylene(h) 37.0 17.2 9.2 9.8 (g)
Polyphenylene oxide 47.0 17.9 3.1 8.5 8.6
Polyphenylene sulfide 38.0 18.7 5.3 3.7 6.7
Polypropylene 30.5 17.7 2.9 1.2 6.2
Polypropylene glycol(h) 32.0 16.5 9.0 7.0 (g)
Polystyrene 34.0 22.8 5.8 4.3 12.7
Polysulfone 42.1 18.5 8.5 7.0 9.4
Polytetrafluoroethylene 19.4 16.2 1.8 3.4 3.9
Polyurethane 37.5(e) 18.8 10.0 8.2 9.8
Polyvinyl acetate 35.3 20.9 11.3 9.7 13.7
Polyvinyl alcohol 37.0 14.7 14.1 14.9 10.5
Polyvinyl butyral 28.0 19.1 9.5 12.2 10.0
Polyvinyl chloride 37.9 16.8 8.9 6.1 8.5
Polyvinyl pyrrolidone(h) 46.0(e) 17.5 8.0 15.0 (g)
Polyvinyl fluoride(f) 32.7 17.4 13.7 11.3 (g)
Polyvinylidene chloride 40.2 19.0 9.6 9.0 5.8
Polyvinylidene fluoride 31.6 17.0 12.1 10.2 4.1

(a) Critical surface tension data is from column 3 of Diversified Enterprises’ Polymer Table 1(4), except as otherwise noted. Hansen Solubility parameters and Ro data are the average of results cited by Hansen(1), except as otherwise noted.
(b) Critical surface tension in mJ/m2 (equivalent to dynes/cm), generally determined by the Zisman method (regression of the cosine of the contact angle), or by the wetting tension method, using solutions of 2-ethoxyethanol and formamide, per ASTM Std. D-2578(3).
(c) MPa1/2 (equivalent to joules/cubic centimeter; 2.0455 x (cal/cc)1/2) @ 20-25oC.
(d) Interaction radius (sphere of influence); units also MPa1/2.
(e) No critical surface tension data available; surface energy data are cited, mostly from contact angle measurements; from column 4 of Diversified Enterprises Polymer Table 1(4), except as otherwise noted.
(f) Solubility parameter data from Yu, et al(5).
(g) No data available.
(h) Solubility parameter data from Abbott(6).
(i) Solubility parameter data from Abbott(7).
(j) Critical surface tension data from Jacobsen, et al(8);

Using the data from Tables II and III and applying formula 1, the Relative Energy Difference between any polymer whose HSP values are known and any dyne level test fluid can be readily calculated. Fortunately, the RED works out to greater than 1.0 for most pairs when considered at the critical surface tension of each polymer.

The most notable exception is polyvinyl alcohol, which has a surface energy of about 36 to 40 dynes/cm. In that range, RED vs. dyne solutions is only about 0.44 to 0.73. Cellophane is another material that could prove difficult to dyne test: its surface energy is about 42 to 48 dynes/cm, with a corresponding RED of 0.63 to 0.96. Of particular note with this material is that the RED is only 0.36 to 0.48 at 38 and 40 dynes/cm, meaning that some degree of surface alteration if testing at these dyne levels is quite likely, even within the brief 2 to 3 second time frame required for the test.

Other polymers that could possibly present problems include polyvinyl acetate (dyne level range of 32 – 36 with REDs of 0.77 to 0.87), poly n-butyl methacrylate (dyne level of about 30, with an RED of 0.87), and nitrocellulose (dyne level range of about 38 to 45, with REDs of 0.80 to 1.15). Given the brief contact time required for dyne testing, it seems unlikely any of these would have a degree of solubility that would have a practical impact, but the possibility needs to be considered.

It is not possible to determine the RED values for the polymers listed that do not have data on their interaction radii Ro. However, of the group only polyoxomethylene and polypropylene oxide appear to have any real potential for solubility with the dyne solutions, as the calculated RA values are 7.4 and 9.8 at 34 and 36 dynes/cm, respectively for the former, and 8.1 at 32 dynes/cm for the latter.

It is interesting that polymethyl methacrylate (acrylic) is not among the polymers cited, as it has always been an axiom that it is a polymer which is especially prone to solubility issues in dyne testing. In fact, according to the data presented by Hansen, the RED for this material works out to 1.17 to 1.42 over a range of 36 to 40 dynes/cm, which reflects its normal critical surface tension range. Possibly some polymerization methods result in significantly different HSPs from the mean, or perhaps some materials which are assumed to have an acrylic surface have actually been coated or otherwise modified. Additionally, if it is an acrylic with unusually low molecular weight, its RO will increase, which could then include dyne solutions within its sphere of influence (as well as making it amenable to a broader range of coatings, inks and adhesives).

Determining potential solubility issues for treated polymer surfaces presents a layer of uncertainty, as corona, flame, or atmospheric plasma all increase the polar component (and likely to a lesser degree the hydrogen bonding component) of HSPs, as well as the polar component of surface energy. There is little information available on this – the only data I could find was for UHMW polyethylene fibers treated by atmospheric plasma. The HSP data presented by Kusano, et al(9) was δD 16.5, δP 15.3, and δH 8.2 hydrogen after plasma treatment. Their baseline (untreated) data was δD 18.0, δP 1.2, and δH 1.4. (By comparison, the average HSP from studies cited by Hansen(1) for untreated PE was δD 16.8, δP 3.8, and δH 3.8. Also of interest is that the Kusano states an Ro of 4.3, compared to 6.6 from the Hansen data; this is likely due to the high molecular weight of the fibers).

Unfortunately, no contact angle measurements or critical surface tension data was reported for the treated fibers in this study, so it is impossible to match these HSP data to that of the dyne solutions that would be used to test the treated material. Additionally, the changes in HSP for UHMW polyethylene likely differ – perhaps substantially – from what would be true for other polymers. To make things even more nebulous, each treatment method (corona, flame, plasma, UV bombardment, etc.) will render a different blend of reactive sites on the surface; treatment dwell time will also affect this; and treatment under non-atmospheric gases in plasma systems can result in dramatically different changes in surface chemistry.

Nevertheless, it seems safe to say that the likelihood of solubility affecting dyne testing results for treated surfaces is probably higher than it is for unmodified polymers: under most scenarios, the major effect of surface treatment is to increase the polar component of surface energy, and the results from Kusano suggest that this would also be broadly true (along with a significant increase in the hydrogen bonding parameter) of its impact on HSPs. A perusal of Tables II and III shows that the dispersion components of the dyne solution HSPs generally match those of the polymers rather closely, so differences or similarities in the polar and hydrogen parameters are the primary drivers of soluble vs. insoluble interactions.

As a rule of thumb, those two components are higher for the dyne solutions than they are for the polymers. Whereas the rate of increase of dyne solution polar parameter is rather steep from 30 to about 38 dynes/cm, the slope decreases progressively as you reach higher dyne levels. The same trend holds true for the hydrogen parameter, at least until you reach the dyne levels (58 and above) that include water as a constituent.

The question boils down to whether or not the polar (and hydrogen) parameter increase is roughly linear as a function of increased dyne level in treated polymers. Only a designed study could answer this question, and it would be a rather daunting challenge, given all the variables involved, including polymer type, treatment method, treatment intensity, etc. This would make for an excellent PhD thesis!

In the absence of actual data, I am willing to go out on a limb on this – a little bit anyway. My guess is that the increase in polar + hydrogen parameters will, in fact, be relatively linear with respect to the amount of surface treatment imparted, at least within a normal treatment range and for treatments that utilize atmospheric gases (that leaves out more sophisticated plasma treatment options).

If this is true, then there is in fact a greater likelihood that solubility issues may affect the results obtained from dyne testing of treated polymers, especially in the range of 42 to 56 dynes/cm. And, as the polar and hydrogen parameters do both continue to climb, albeit at a reduced rate, this may be the explanation for why spurious wetting is sometimes observed when testing is mistakenly continued at levels above the critical surface tension (as indicated by a beading time of 2 seconds): That extra boost in these solubility parameter components may be the critical difference between insolubility and solubility of the polymer in the dyne solution. The latter scenario would lead to artificially easy wetting.

References:

(1) C.M. Hansen, Hansen Solubility Parameters: A User’s Handbook, 2nd Ed., CRC Press, Jul 2007. 

(2) S. Abbott, “The HSP sphere,” https://www.hansen-solubility.com/HSP-science/sphere.php.

(3) no author cited, “ASTM D2578-17: Standard test method for wetting tension of polyethylene and polypropylene films,” ASTM, 2017.

(4) R.E. Smith, “Critical surface tension, surface free energy, contact angles with water, and Hansen Solubility Parameters for various polymers,” Diversified Enterprises, http://www.accudynetest.com/polytable_01.html.

(5) W. Yu and W, Hou, “Correlations of surface free energy and solubility parameters for solid substances,” J. Colloid and Interface Science, 544, 8-13.

(6) S. Abbott, “Make your own designer solutions,” https://www.hansen-solubility.com/HSP-science/solvent-blends.php.

(7) S. Abbott, “HSP examples: poly lactic acid (PLA),” https://www.hansen-solubility.com/HSP-examples/pla.php.

(8) J. Jacobsen, M. Keif, X. Rong, J. Singh, and K. Vorst, “Flexography printing performance of PLA film,“ https://digitalcommons.calpoly.edu/cgi/viewcontent.cgi?article=1036&context=it_fac.

(9) Y. Kusano, S. Teodoru, and C.M. Hansen, “The physical and chemical properties of plasma treated ultra-high-molecular-weight polyethylene fibers,” Surface and Coatings Technology, 205, 2793-2798.