GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

IGUs to Replace Divided-Lite Windows

gordy_b | Posted in Energy Efficiency and Durability on

This true divided lite 2x-paned window (pic attached) is on the east side of a home in climate zone 5a.  Today, while stripping to re-putty, the contractor pointed out foggy-ish panes and broke a pane, leading to the idea of replacing them all.  I wish to maximize the energy efficiency of this large surface to the extent possible only replacing IGUs.  But, not sure a couple hours of reading prepared me to pull the trigger on this tomorrow.  Please comment on my tentative plan… First, general specs as the only way I know to get these IGUs is the contractor makes them as 2x-pane Argon IGUs.  So probably no NFRC label. The old ones are 3/8″ deep; he proposes the new ones be 1/2″ (am limited to 1/2″ due to munton depth and the necessary putty).  There are no 1/2″ 3x pane windows, correct?  Making new 2x-pane IGUs 0.125″ (=3 mm) deeper will decrease U, right? Spacers should be anything other than aluminum, preferably foam, right? The low-E coating should be on surface 3 (2cd surface from the interior on 2x-pane windows), correct? SHGC should be 0.42 or higher as this is the Energy Star (ES) northern climate (see ES window specs from ’20, attached).  The shop-made IGUs likely won’t be near ES U values and in the chart, as U-value gets worse SHGC increases so I guessed SHGC above 0.42 would be best.  Reasonble? Second, the contractor installs windows very well.  But they likely don’t have the manufacturing options to provide ES or better specs (I know the whole-window specs will be worse than those of the IGUs but I need something to aim for).  Googling high performance IGUs returns results on full framed windows, but little about high performance IGU replacements for divided lites.  Isn’t there a way to have Alpen or another high performance company make these to high specs then have the contractor install them?  What am I missing? Thank you very much, Gordon

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. GBA Editor
    Kiley Jacques | | #1

    I'll give your a question a bump and share this source I found for simulated divided lite windows with IGUs: Agnora. I admit I am unfamiliar with this manufacturer but it might be worth investigating their products and services.

  2. gordy_b | | #2

    Thanks Kiley, but I'm not ready to give up on the idea of rehabbing an existing multi-pane window for efficiency. Maybe others will write something like "limited to 1/2" full IGU depth you'll never begin to approach that surface area efficiently retaining heat, even with insulating shades etc." That would nudge me toward the simulated multipane window down the road. Hope kills!

    Thanks again!

  3. [email protected] | | #3

    Replies to your questions:

    "The old ones are 3/8″ deep; he proposes the new ones be 1/2″ (am limited to 1/2″ due to munton depth and the necessary putty). There are no 1/2″ 3x pane windows, correct?"

    When you say 3/8" deep are you talking about the full thickness of the IGU or the width of the airspace between the lites? In either case there will be an improvement going wider airspace. If you are talking airspace width then the improvement will be more pronounced.

    There are 3 pane IGU's with two 1/2" airspaces, if that is what you are asking. On the other hand trying to build a 3 pane IGU with an overall thickness of 1/2" would be ludicrous.

    "Making new 2x-pane IGUs 0.125″ (=3 mm) deeper will decrease U, right?"

    Yes, but how much improvement again depends upon if you are talking about overall IGU or simply the airspace between the glass.

    "Spacers should be anything other than aluminum, preferably foam, right?"

    Aluminum is not your best option. Foam is fine but more concerned about what foam used and how the contractor puts them together.

    "The low-E coating should be on surface 3 (2cd surface from the interior on 2x-pane windows), correct?"

    Surface 1 is the outside exposed glass face and surface 4 is the inside exposed glass face. LowE coatings are generally applied to surface 2 unless you want higher SHGC then specific high solar gain coatings can be applied to surface 3. Any coating applied to surface 2 will lower SHGC and applied to surface 3 will raise SHGC. U factor is unaffected by surface 2 or 3 application.

    Also, what LowE is he using and how is it stored, handled, cut, installed in his shop? If it's a soft coat then they are very fragile and easily damaged during IGU manufacturing. If he uses a hard coat, then which one becomes important because of performance differences.

    "SHGC should be 0.42 or higher as this is the Energy Star (ES) northern climate (see ES window specs from ’20, attached). The shop-made IGUs likely won’t be near ES U values and in the chart, as U-value gets worse SHGC increases so I guessed SHGC above 0.42 would be best. Reasonble?"

    Unless you are specifically looking for a high solar heat gain application (only facing south and only with proper shading in summer for that window), then go for the best U factor that he offers. Unfortunately as a shop-made IGU no real information for you to use to choose.

    "Second, the contractor installs windows very well. But they likely don’t have the manufacturing options to provide ES or better specs (I know the whole-window specs will be worse than those of the IGUs but I need something to aim for). Googling high performance IGUs returns results on full framed windows, but little about high performance IGU replacements for divided lites. Isn’t there a way to have Alpen or another high performance company make these to high specs then have the contractor install them? What am I missing?"

    When looking at U factor, SHGC, etc., you should be comparing full window values and not simply glass values. You are proposing adding 25 individual IGU's to an existing window so that you want to maximize IGU performance, and while that is admirable and a great goal, reality it that no matter what you do to try to improve that window, the energy performance isn't going to be very good because of the nature of the beast.

    The most efficient part of any IGU is called "center-of-glass" (CoG). CoG is all glass area that is not within 2-1/2" of the edge of the glass. Doesn't matter if the IGU is 12x12 or 120x120, the CoG thermal performance will be the same as long as the two IGU's are otherwise identical.

    The outer edge 2-1/2" (EoG) will be colder or hotter than the CoG. How much depends on spacer system used as well as LowE coating and how the IGU is glazed into the sash.

    If you are building a 120x120 window, then EoG has relatively little effect on overall IGU energy performance, but if you are building a 12x12 window then that EoG lower performance 2-1/2" has a great deal of effect on overall window performance since it's a much higher percentage of total glass. Multiply that by 25 separate smaller IGU's with each and every muntin bar as an energy performance loss, plus and even more importantly, the potential for air leakage along the edge from every single IGU installed - 100 separate edges for TDL versus 4 for an SDL - is a nightmare for overall energy performance no matter the quality of the IGU installed.

    Finally from a long term performance perspective, a smaller IGU introduces a greater risk for seal failure than a larger one, so multiply that greater risk by 25.

    Despite best intentions, while you can improve the performance by upgrading the individual IGU's, ultimately and unfortunately it's still going to be overall pretty bad.

    I can't imagine that replacing 25 smaller IGUs is less expensive than a full replacement with a fixed SDL, of that's a concern, but if it is then that would scare the heck out of me.

  4. gordy_b | | #4

    Thanks Oberon,

    Sounds like the Edge-of-Glass (EoG) effect on a 25-pane window makes getting this window even OK from an efficiency perspective impossible - "while you can improve the performance by upgrading the individual IGU's, ultimately and unfortunately it's still going to be overall pretty bad." I haven't seen out-of-the-box SDLs that look like this. So it'd have to be custom, or hire designer/architect to choose an out-of-the-box SDL. So let's make this one as "less bad" as possible...

    The total depth (two panes + gap) of the old IGUs is 3/8" and this can be increased 3 mm to 1/2". Your reply and the attached image showing 2-pane Argon-filled IGUs improve markedly with minor depth increases around 0.5" of total depth ( ) suggest a 3 mm improvement over 3/8" is worthwhile.

    Spacers... Contractor will outsource the IGUs to what I think is an actual glass company that can do foam spacers. So, the foam should be well installed, should modestly thermally improve the window, and maybe decrease the rampant condensation. Any down sides to foam spacers??? Are they structurally stable over time, do they let the Argon out quickly...???

    Low-E type... The contractor says the type of Low E coating is Low E2. Is there a better coating for zone 5a east-facing windows???

    Low-E2 on surface 2 or 3... "high solar heat gain application (only facing south and only with proper shading in summer for that window)" (Oberon). Energy Star windows for the northern area increase the SHGC as U-value increases (attached above). These 1/2" IGUs with much EoG effect will necessarily have a higher U-value. Wouldn't it be in keeping with the Energy Star idea to increase SHGC by having the low-E on surface 3 even though the window is east-facing??? I'm not opposed to Low-E on surface 2, just checking.


  5. gordy_b | | #5

    FYI - my answer to Oberon's "What Low-E are they using?" was Low-E2 because I was told it would be like Marvin Windows' Low E2 designation. I just read on the Marvin Window site that the "Low-E2" offered by the glazing shop means "a double layer of silver on an inside surface of IG glass... [offering] low U-Factor [and] medium solar heat gain" ( ). So this is NOT the meaning of Low-E2 some use to indicate two surfaces have Low-E coatings.

  6. gordy_b | | #6

    The Less-Bad True Divided Lite Rehab/Upgrade

    With 25 IGUs the Edge of Glass effect means this will never be even an OK window energy-efficiency-wise. But would the following make it about as less bad as it can be while minimizing condensation potential? I'm using Cardinal's 1/4" air space info because that's as small as they have and as mine must be 1/2" full thickness, I don't know what the air space would be but it would be minimal...

    All Cardinal's following have U-factors of 0.25 ( , Fig. 2-4 ). And they all ( ibid, Fig. 2-35) resist condensation up to 37% RH at 0 degrees F...

    Argon-filled LoE-366® (#2) / LoE-i89® (#4)
    Argon-filled Quad LoE-452+™ (#2) / LoE-i89® (#4) AND
    Argon-filled LoE-340® (#2) / LoE-i89® (#4)

    Of course the whole window with all the edge effect would not approach 0.25 U BUT...

    In contrast, the contractor has suggested something like an argon-filled Marvin Low E2 (two layers of silver on surface 2). Cardinal's argon Low-E 272 is two layers of silver on surface 2 and the 1/4" gap version has a U-factor of 0.32. Presumably what the contractor is suggesting would be U ~ 0.32. Again, my window with all that edge effect would not approach 0.32 U.

    BUT, wouldn't using one of the Cardinal combinations with the room-side Low-E so that all 25 IGUs are roughly 0.07 U-factor less than the Low-E-only-on-surface-2 IGUs make for a much less bad window (no matter what the whole window U turns out to be)???


  7. [email protected] | | #7

    LowE 2 coating would be the one that I would have selected for your environment as well.
    Marvin uses Cardinal's LoE^2-272, but there is also LoE^2-270 that has similar performance characteristics. Either would be acceptable for your environment.

    You mention Cardinal's 366, 340, and 452 as possible options. 366 and 340 are triple-silver layer coatings, and 452 is a quad-silver layer coating. These are all low solar heat gain coatings.
    However the condensation temperatures that you quoted from the Cardinal chart are center-of-glass and are not reflective of edge temperature. The edge is colder than the center.

    You also mention I89 as a possible option along with these coatings for U factor improvement, but because i89 improves U factor performance by reflecting radiant heat back into the room, radiant heat that would have been partially absorbed by the glass if there was no surface 4 coating, i89 actually LOWERS condensation resistance because reflecting heat back into the room results in colder glass.

    Any chance the contractor would share who is making the IGU's? Although they are using Cardinal information per coatings, it's certainly possible that they are actually using different glass/coatings but using Cardinal labels because they are the most common and most recognizable. I am guessing that they probably use superspacer foam spacer for their IGU's. It's pretty common and very well regarded by folks who use it.

    The only really relevant, to your situation, part of the LBNL chart that you posted is the solid blue line, and once again that only reflects CoG and not the edge. This is because, in part, that the edge is directly affected by the size of the IGU. Less effect for larger IGU's more effect for smaller.

    Your current IGU's, based on 3/8" overall thickness, are likely to be
    2.2mm glass / 1/4" air space / 2.2 glass. 2.2mm is also called single strength glass.

    At 1/2" there are two possibilities - 2.2mm glass with 3/8" airspace or 3.0mm glass with 1/4" air space. 3.0mm is also called double strength glass. Given the option, the 3/8" air space is better despite using single strength glass.

  8. gordy_b | | #8

    Oberon, thanks for sticking with me.

    YES, the contractor is getting me in touch with the glazing shop (which uses Cardinal). I'll post the options they offer when I find out (maybe they don't have 366 and it's two months back-ordered). Unless you're highly motivated, maybe it doesn't make sense for anyone to reply until I find which Cardinal options this glazing shop offers.

    I am convinced decreasing SHGC and increasing U as Oberon suggested way above. To that end we're set on a Low-E of some sort on surface 2.

    The suggestions of Cardinal's 366, 340, and 452, each (#2) with i89 (#4), were based on reducing U-factor and SHGC as much as possible and staying well above Cardinal's suggestion of 30% RH at outside = 10 F (attached) (average coldest across last January in this suburb of Chicago was 13 F). I hope given the CoG on the combos above at 1/4" gap start to gather condensation above 37% RH at outside = 0 F then the edges might still be condensation-free at the 30% RH I'll need in January's roughly outside = 10 F.

    The room-side i89 question I suppose is the same as above. If CoG of 366, 340, and 452, each with i89 gathers condensation above 37% RH at 0 F can we say the edges will probably be condensation-free at the 30% RH at 10 F needed in much of January??? Cardinal is using their own steel warm spacers in these charts. While I have no direct comparison to Cardinal's spacers, the attached graphic ( ) suggests the SuperSpacer is probably warmer than anything. Do I gain a condensation safety buffer if I can use the SuperSpacer???

    The LBNL chart was to convince myself U-factor changes sharply around 0.5" of IGU depth. No IGU in the window has broken (other than when someone was working on it) and that was probably the 2.2 mm glass. So, other than jinxing myself by saying this out loud, I'm certainly comfortable with the same thickness glass pieces and the better 3/8" gap - U-factor would improve, condensation would decrease.

    Thanks, I'll be back after being in contact with the glazing shop.

  9. gordy_b | | #9

    Oberon (or others knowledgeable about 2x-pane stack-ups),

    Have I found another stack-up option???

    2.2mm-3/8" argon-2.2 mm or 3 mm-1/4"-3 mm were the options to reach the 1/2" IGU thickness limit we need to leave for enough room for puttying to the muntons. Of the two, the 2.2 mm-3/8"-2.2 mm setup was thermally preferred, which I figure is 0.55" = 35/64." The contractor agreed to this if we can source 2.2 mm and agreed to the fallback of 3 mm-1/4" argon-3mm.

    But, I think that 3 mm-5/16" argon-3mm sums to 0.55" = 35/64" just like the preferred stack-up above. Figure 2-5 of the Cardinal Tech Guide shows increasing from a 1/4" gap to5/16" gap improves the U-factor from 0.26 to 0.23. Any reason (like I did the math wrong) that if we must go with 3 mm glass that we shouldn't go with a 5/16" argon gap to get this roughly 10+% U-factor boost???


  10. Trevor_Lambert | | #10

    Not only will making that 25 piece window energy efficient be impossible, it looks horrible. Probably the ugliest window I've ever seen. It also doesn't make much sense to worry about visible light transmittance of the glass when half of the light and view is blocked by mullions. Just replace it with a double pane window and enjoy the improved view and appearance.

  11. [email protected] | | #11

    Looking at the picture of your window I get the impression that it's not flat to the wall but that it has a very slight bow out. If so is that why you can't find anything quite the same in a SDL version and you are attempting to upgrade the original?

    Guessing that the existing IGU's are roughly 18"x9" based on the picture. Going by that assumption each panel would be 162 sq inches or 1.125 sqft. Keeping in mind that the U factor numbers that you are quoting are center-of-glass, and that center-of-glass is everything inside of the outer 2.5", each panel consists of 52 sq inches center-of-glass and 110 sq inches edge-of-glass, or roughly 32% of each panel is CoG and 68% is EoG. So the very best U factor that you can squeeze out of your glass make-up will only apply to 32% of the individual IGU, while for 68% of the total area the U factor numbers doesn't apply.

    I have no idea who will be making your IGU's but I am guessing that you won't be getting them for anything less than about $40 each or about $2000 for the bunch? I would be surprised If they come in less than that but less surprised if they come in higher. And then labor to to install they panels has to be significant, especially if the installed has to router each individual opening to fit the 1/2" versus 3/8" IGU. And for a finished product that isn't really going to be much better than what it's replacing no matter what you do. There simply is no way to avoid 100 separate edges and make them efficient.

    My strongest recommendation is to replace that window, but since you asked what is the best possible scenario based on knowing that performance won't be very good, that's what I have been going with. But in good conscience I am not comfortable with any scenario that doesn't replace that window.

    Moving on.....

    3.0 / 1/4" air / 3.0 = .486" U.26
    2.2 / 5/16" /2.2 = .486" u.23
    In this scenario 5/16" is the obvious better choice.

    3.0 / 5/16" / 3.0 = .549" U.23
    2.2 / 3/8" / 2.2 = .548" U.21
    In this scenario 3/8" is better than 5/16" IF the extra .05" isn't a problem.
    Again given that the U factor is only valid for about 32% of the total glass area.

    The one place where you could gain some level of improvement would be radiant heat rejection in summer and heat retention in winter due to using a better LoE coating than what's currently in the windows (not knowing if they are LowE or if so what LowE), but that improvement will likely be offset by the general inefficiency of the window in general.

    Assuming the window is flat to the wall and doesn't have a bow, have you discussed the idea of carefully removing the current muntin bars and trimming them down so that they could be applied to a new flat SDL IGU using VHB tape? VHB is what window companies use to apply SDL muntins to the glass surface. VHB is permanent and once applied you won't be getting them off in the future.

    Even with a slight curve, it's possible to attach to a new SDL, just a bit more work setting it up.

  12. [email protected] | | #12

    This is how Cardinal builds an IGU.
    The opening building is Cardinal's IG Product Development Lab in Minneapolis, as are the portions of the video showing testing and development. The IGU assembly portions of the video are from Cardinal's Spring Green WI IG plant (down the street from the AG shown in video), only a couple hours from Chicago.

  13. gordy_b | | #13

    Oberon - that video is amazing.

    I'm currently trying to source Super Spacers to reduce the EoG effect as much as can be done for this window.

    But, part of me wonders if Cardinal's warm edge spacers (almost as good as Super Spacers) + their very precise operation is the way to go relative to reducing future fogging - I don't know how good local window shops are at getting the seals right and keeping the argon in (and the window contractor says Cardinal will be much more expensive). Thoughts on local vs Cardinal making the IGUs anyone???

  14. [email protected] | | #14

    Compare to Cardinal IG video:

    Typical small shop superspacer application method:

    Automated superspacer application:

    Cardinal's Endur stainless steel spacer is .8°f lower temperature at the edge than the best superspacer variant (7 variations last time I looked), while Cardinal's XL is about 2.5°F lower temperature than the best superspacer variant. Endur is actually warmer than at least 2 of the non metallic supsrspacer variants with significantly longer life cycle expectancy over any non metallic spacer system and with better long term argon retention performance.

    All that said, Cardinal generally only supplies directly to OEM window manufacturers, but I have heard of possible exceptions now and then from some Cardinal plants if business volume permits.

    If you are concerned primarily about seal failure (fogging inside the IGU), Cardinal currently has over 500,000,000 IGU's in the field under warranty and over 200,000,000 units in the field past warranty that are performing with a seal failure rate of .2% at 20 years. Actually, the XL was introduced 30 years ago and since introduction the field failure rate is still .2% even at 30 years. (to confirm, that's 2/10's percent, not 2 percent).

    If you are worried primarily about moisture or frost on the number 4 glass surface inside your home, that's more a function of the TDL's and not much you can really do about that without going to a SDL unfortunately.

    Cardinal technology overview:

  15. gordy_b | | #15

    Thanks Oberon, the shop method looks more precise than I had feared.

    If I hadn't looked into this, the contractor would have installed either no coating or LoE 270 only with aluminum spacers. I think either a LoE 270 with i89 and Super Spacers made by a shop OR LoE 270 with i89 and Endur (or XL) spacers made by Cardinal will be much better (though nowhere as good as a new window I know). One might have slightly better spacer insulation, one might retain argon longer but, given this is a retrofit I'll take whichever of these improvements I can get.

    So, super spacers and the shop will be the first option seeing Cardinal usually only sells to OEMs. I'll try Cardinal itself if for some reason the first option doesn't pan out.

    1. [email protected] | | #16

      I would be very wary of using i89 in your situation. Using i89 will lower glass temperature and can result in increased chance of window condensation when it's cold outside, especially edge condensation.
      In your situation I suspect that using i89 will almost certainly result in wet windows in winter unless your inside humidity levels are rock bottom when it's cold outside.

  16. gordy_b | | #17

    A few items I've neglected to address...

    There might be hope that the CoG effect will be a larger proportion of the whole window's surface. I measured an IGU and came up with 46% of each IGU being CoG (I did not calculate the areas, I just had SketchUp tell me the area of each section, attached). So while it still won't make a great window, it's better than the 32% you had to guestimate above because I had not provided measurements.

    Heat retention in winter compared to coatings on the current IGUs... The current IGUs are mostly from the '50s with some having been replaced in the '60s and '70s. They have no coatings that I am aware of. LoE2 270 + i89 should be a large improvement relative to these old IGUs.

    Carefully carving up the muntin frame then taping back on... I'm not sure if the curve (attached) is too much for this. But this far into the project I'd better stick with the retrofit-what's-there-using-the-current-muntin-framework plan even if deconstructing the muntin framework and taping it back onto new windows would work with this curve.

    Thanks again, I'll be back after I secure spacer options.

  17. gordy_b | | #18

    WoW! Took a month for my contractor's preferred window shop to say they can't make IGUs with the narrower Super Spacer needed for these old muntons. My contractor then asked other IGU-shops and none of them could. So we're going with Cardinal IGUs. The specs follow, but first...

    I respect Oberon's opinions, changed at least one spec based on them, and KNOW I'M TAKING A BET on the following... Pressured from the contractor to order the IGUs to get them in before winter I specified i89 without checking the following reasoning with the GBA community...

    The muntons' interior have very little damage after 70 years of dual-pane, thin gas space, no argon, aluminum spacer windows, and some condensation. The new IGUs will have deeper gas space, be argon-filled, and have Cardinal's Endur IG warm edge spacer. While some of these might not warm the EoG much, I'd hope they'd all contribute. So unless i89 cools surface 4 EoG more than these improvements warm it, the condensation won't be worse than before. Given the lack of munton damage, this might be OK. And, if the i89 doesn't cool as much as the improvements warm, then 4 will be that much warmer. We'll see! :-(

    So... Neat+ on surface 1 (in hopes it actually works to keep the window cleaner), 2.2 mm glass, low-E 270 on surface 2, 3/8" gap filled with argon, Endur IG warm edge spacer, 2.2 mm glass, and i89 on surface 4. I know it's not even a good window. But, maybe it's close to the best it can be while keeping the family member who is attached to it happy.

    Thanks all, especially Oberon! I might have follow-up ?s but will likely start new thread(s).

    1. gordy_b | | #19

      Window Folk (especially Oberon),

      Update on Bow Window Restoration...

      Thanks for all the help, I think the specs turned out about as good as one could expect for a retrofit. The following stack up was implemented last Fall with one possibly major addition... Neat+ on surface 1, 2.2 mm glass, low-E 270 on surface 2, 3/8" gap filled with argon, Endur IG warm edge spacer, 2.2 mm glass, and i89 on surface 4. The ADDITION was trim strips at the 90 angle between the old wood frame and the new IGUs. One family member liked the look and the other thought the strips might decrease the Edge of Glass coldness. Good news... don't remember if last winter was less cold than most, maybe the trim strips helped insulate the edge, anyway there were only one or two incidents of interior condensation (and they might have been on broken new IGUs that have since been replaced). I'm not recommending these specs to anyone, but in this instance, so far so good.

  18. Deleted | | #20


Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |