Durable and Economical Technology for Bituminous Surfacing of Rural Roads
The Pradhan Mantri Grameen Sadak Yojna (PMGSY) has been one of the most successful flagship programmes launched by former Prime Minister Atal Bihari Vajpayee in 2000 for constructing a wide network of rural roads across India. During the last 17 years, a total of 5,29,975 km of PMGSY roads have been constructed connecting villages of over 500 people, which is a commendable task!
Two technologies (surface dressing and premix carpet) are specified in India’s ‘Rural Roads Manual’ for bituminous surfacing of rural roads such as those under PMGSY. Either of these two technologies are permitted. Both are also standardized by the Indian Roads Congress (IRC).
Surface dressing (also called chip seal) consists of spraying a thick film of bitumen over the compacted stone base layer (called water bound macadam) with the help of a bitumen truck tank distributor at a specified rate. This is followed by spreading stone chips (aggregate) at a specified rate by a mechanized chip spreader (available from many manufacturers in Gujarat and costing less than Rs. 2.5 lakhs). Stone chips are then compacted with a road roller so as to embed/fix them (about 70%) into the sprayed thick bitumen layer. This completes the surface dressing operation as illustrated in Figure 1.
The premix carpet (PMC) simply consists of single size (about 12 mm) stone chips mixed with 3.5% bitumen by weight at a hot mix plant. PMC is highly open graded because the mix does not contain any fine aggregate (sand). This mix is laid in 20 mm thickness manually or with a paver. A sand seal coat is provided at the top.
The objective of this paper is to discuss these two technologies in detail and recommend the one which is better both in terms of durability and economics.
History of Bituminous Surfacing
When the author was serving as a highway engineer in the Rajasthan PWD during the early 1960s, it was very common to use bituminous surface dressing (SD) or chip sealing on most types of roads. Surface dressing was very effective in waterproofing the water-bound macadam (WBM) roads because of heavy bitumen application rate followed by chip application. Surface dressing was scheduled once in 3 or 4 years on all roads. Very few potholes dotted the roads at that time, and traffic volumes were generally less. Road construction was largely manual and hardly mechanized. Bitumen for surface dressing was applied with perforated tin cans. Spreading the surface aggregate (chips) by hand was an art learnt through practice, usually by swirling the basket containing the aggregate.
As is often usual with surface dressing, chips were dislocated and became loose if the treated road was opened too soon to traffic, or slow speeds were not maintained just after construction. The finished road surface was not black and, therefore, not too appealing to the public.
To overcome these perceived “problems”, the premix carpet (PMC) was introduced with the IRC publishing its specification for the first time in 1962. As mentioned earlier, road construction was still manual. Single size chips (nominal size 12 mm) were either broken by labourers by hand or obtained from stone crusher plants (if available nearby). Hot bitumen was applied
as tack coat through perforated tin cans. The mix containing almost single size aggregate (11.2 mm to 13.2 mm) could easily be coated with about 3-3.5% bitumen either by hand on flat pans placed over wooden log fire; or small drums rotated by hand; or small portable mixing plants. Under such circumstances, graded aggregate could not be used.
The mix was taken in hand carts and spread over tack coated road surface using hand rakes. After rolling, the road surface appeared “shining” black, with no loose stones and impressive looking to the public, unlike surface dressing. It was realized that PMC was highly permeable to rainwater due to single size aggregate being used in the mix. Therefore, the use of sand seal coat was warranted to seal the surface of the open graded mix. Sand was mixed with about 7 percent bitumen, applied on the open surface, and rolled. With the advent of the PMC, surface dressing started to die across India and is almost non-existent in many states such as Rajasthan. It is ironical that surface dressing is still being used extensively and successfully on low to medium-trafficked roads in developed countries such as USA, Australia, New Zealand, and South Africa. Some Indian engineers argue that surface dressing is successful in those countries because the construction (bitumen application and chip spreading) is mechanized there. However, the engineers can mandate bitumen distributors (already available for tack coat work) and mechanized chip spreaders (being manufactured in Gujarat). A typical chip spreader is shown in Figure 2.
The PMC has probably served India well for over 50 years especially during the time mechanization was almost not there. However, due to significant increase in vehicular traffic, and PMC’s inherent watertrapping characteristics, its service life has decreased significantly in recent years. Time has now come to think out of the box and consider surface dressing in lieu of PMC for low to medium-trafficked roads, because it is highly economical as well as highly effective in water-proofing the road pavement (as discussed later).
Premix Carpet (PMC)
The undesirable water-trapping characteristic of the PMC, which causes potholes due to increased hydraulic pressure under traffic, is discussed below: To keep things in perspective, let’s compare PMC with open graded asphalt friction course (OGFC), which is used in developed countries primarily for road safety. Although, OGFC is not used in India, experience with OGFC is applicable to PMC used in India in certain aspects. Both are highly water permeable (porous) mixes and are placed 20mm thick. The OGFC is placed on dense bituminous concrete (similar to BC Grading 2) to provide a skidresistant wearing surface during rainfall or when the pavement is wet. The rainwater penetrates the open surface of the OGFC; goes to its bottom; then flows within 20mm thick OGFC towards the shoulders; and then exits from the exposed edge of the OGFC onto the shoulders. Since there is no rainwater on the surface of OGFC, there is no hydroplaning or skidding of motor vehicles on its surface. OGFC is highly permeable to water since it has over 18% air voids. OGFC is durable despite high air voids because it has about 6 percent polymer modified bitumen content, which provides a thick bitumen film around the aggregate particles.
The premix carpet (PMC) on the other hand, is substantially more open graded and more porous than the OGFC because the former uses very coarse aggregate (nominal size of 11.2 to 13.2mm) and no fine aggregate. Its air void content is estimated to be over 25%. Although, a sand seal coat is provided on the surface of the PMC, it is not completely effective in making the PMC water-proof at the surface. Even if there is a small patch where the PMC has lost its sand seal, the surface water on the road can penetrate it at that spot, flow sidewards like in OGFC, and flood the entire PMC below the sand seal (Figure 3). The hydraulic pressure induced by traffic in the water trapped within the PMC below the seal coat is likely to cause stripping within the PMC and the underlying bituminous course. If the underlying course is water bound macadam (WBM) or water mixed macadam (WMM), it would get saturated and lose its strength, especially if it contains some plastic material.
The surface water permeability of an inservice PMC was determined recently with a grease ring method. Although it is a simple, crude, falling head water permeability test, it does give some relative permeability values. A ring of about 225mm in diameter and about 25mm high is made on the road surface to be tested using heavy grease. Putty can also be used in lieu of heavy grease. The ring is filled with water upto a depth of 12.5mm and a timer is started. Time taken by the water to penetrate and disappear from the road surface is measured in seconds as a measure of relative water permeability.
The first test (Fig. 4) was made on PMC without any seal coat. It was not even possible to fill the ring with water because it was penetrating the PMC as fast as it was poured. On filling rapidly, water penetrated fully in about 5 seconds. The second test (Fig. 5) was made on PMC with moderate amount of sand seal coat. The measured field permeability was 105 seconds. The third test (Fig. 6) was made on PMC with adequate amount of sand seal coat. The measured field permeability was 545 seconds, which is still considered very high. It is not uncommon to see nonuniform application of sand seal coat on PMC because it is usually spread manually. It is a matter of great concern. During a similar test on BC wearing course, water remained at 12.5mm level for hours, therefore, the field water permeability was almost zero (Fig. 7).
More field permeability test data has been obtained by students at Rasta, Bangalore; IIT Guwahati; MNIT, Jaipur; and Kautilya College of Engineering, Jaipur. A steel ring (Figure 8) rather than grease or putty ring was used to expedite testing. Unlike BC, the test data has indicated a very wide range of water permeability from very high (60 seconds) to almost none. As mentioned earlier, it appears to be a function of variability in sand seal coat application in terms of its quality, quantity and mode of application. Practically, it is not possible to apply consistent and adequate amount of sand seal coat throughout a PMC project. Recent detailed investigations by IIT Guwahati have also shown PMC to be highly permeable to water.
It is quite evident from the preceding field experiments that generally the PMC with sand seal coat would easily take in and trap water during rains in many cases. Once the PMC is saturated with water, the hydraulic pressure resulting from traffic above can
loosen up the sand seal in other areas of the PMC. This phenomenon has been observed on Jaipur streets (Figure 9). As already mentioned, the hydraulic pressure also causes stripping in the PMC as well as in the underlying bituminous courses, if any. That is why; PMC deteriorates rather rapidly during monsoons, especially in towns and cities where streets usually get flooded. The average life of PMC in Jaipur is about 1-2 years. Its bitumen content is about 3.5 percent.
Obviously, there are cases where PMC with good, uniform sand seal coat and/ or very dry climate has endured well. However, fundamentally, the question is why to place a highly porous bituminous mix like PMC in the first place and then try to seal it. There is no available data as to what depth, if any, the estimated 6mm thick sand seal coat really penetrates the 20mm thick PMC when rolled.
PMC is used only in India; it is not used in developed countries. There are numerous other questions related to PMC which need to be answered: total air voids in PMC; absolute volume of sand seal coat; unfilled voids in PMC; depth of sand seal penetration in PMC; its performance and durability, etc. It is surprising to note as to why no such research was conducted in India during the last 60 years to answer these legitimate questions. On the other hand, hundreds of research papers have been published across the world on surface dressing in terms of its rational design, construction and performance. There is hardly any research in India where PMC is used, especially on its structure, volumetric, performance and durability. It appears there is only a “conventional wisdom” that PMC does work and is “good” for India and, therefore, there is no need for any research on it.
No published data on the average life of PMC in India could be found either. Some PMGSY engineers revealed its average service life to be 2 years without significant distress such as ravelling and potholes. Above all, the cost of PMC is about three times the cost of surface dressing. This is not acceptable.
If the PMC is a panacea for low to medium trafficked roads in India, why is this technology not used in developed countries? However, that would require sound engineering justification, which is almost non-existent and hard to come by in case of PMC.
Time has come now to ban PMC altogether because its continued use cannot be justified technically as well as economically, be it city streets, low volume roads (such as PMGSY), or medium to high volume roads.
The alternative for PMC in India is surface dressing, which, as described earlier, has the following advantages:
• It is very cheap at ₹2.62 lakhs per km compared to PMC’s ₹7.88 lakhs per km.
• Due to thick sprayed film of bitumen, it is highly effective in waterproofing the road. Water is the biggest enemy of bituminous road and should be kept away from the road surface.
• Minimizes oxidation of bitumen because it is present as a thick film and stone chips provide protection from sun rays
• Environment-friendly because chips need not be heated
• Contractors cannot use lower bitumen application rate because after construction the stone chips would be lost with the movement of traffic
• Indian Roads Congress has an excellent standard for designing and ensuring good quality surface dressing
• Surface dressing is used successfully in both developing and developed countries (such as USA, Europe, and Australia) for blacktopping low and medium trafficked roads. No other country besides India uses premix carpet
• Hundreds of research papers have been published on surface dressing across the world to fine tune this technology. (The author has published two papers in international journals).
For low to medium-trafficked roads where PMC is used right now, it is recommended to use single or double surface treatment. If ‘black” road surface is desirable for surface dressing to “impress” motoring public as
well as minimize chip loss, precoated chips can be used. It should be noted all these alternatives are much cheaper than the PMC (as shown in Table 1). Note that the cost of single coat surface dressing is only 1/3 of the cost of PMC. The cost of the same with precoated chips increases to only 40% of the cost of PMC. It is not understood as to why it cannot be used on low volume roads such as PMGSY, which would save India thousands of crores of rupees every year. Even if double surface dressing with precoated chips is used, its cost is only three-fourth the cost of PMC. Obviously, the highway agencies have to mandate the use of mechanized bitumen distributor and chip spreader, which are already available in India, to ensure the functional success of surface dressing.
Bituminous Concrete in lieu of Premix Carpet
For medium to heavy-trafficked roads and city roads, use BC Grading 2 in lieu of the PMC. Although it is permissible to lay BC Grading 2 in 25-40 mm depth according to IRC:111- 2009, it is preferable to use 40mm depth to ensure adequate compaction during construction (thin lifts cool rapidly). Although the initial cost of 40mm BC Grading 2 is about 50% more than the cost of 20mm PMC, BC Grading 2 is actually 24.1% cheaper than the PMC based on lifecycle cost analysis (LCCA) given in the Annexure. This is a very conservative analysis in that the remaining service life, salvage value, maintenance expenses, and user operating costs were not considered, which all favour BC. Therefore, savings will be much more than 24.1 percent. More importantly, BC Grading 2 provides significant structural strength to the road pavement for future traffic growth, whereas PMC has almost zero structural strength. If there is some reluctance to use surface dressing in lieu of the PMC on rural roads, 40mm BC Grading 2 can be placed directly on water bound macadam (WBM) or water mixed macadam (WMM), as practiced in developed countries.
Enormous Cost Savings
Cost savings of over ₹22,000 crores could have resulted if surface dressing was used in lieu of PMC since the inception of the Pradhan Mantri Grameen Sadak Yojna (PMGSY) in 2000. In addition, if the changeover is made now, cost savings of over ₹4,400 crores can be realized every year in resurfacing already constructed rural roads. This big saving can be used in building and upgrading more lengths of rural roads in India.
The preceding cost savings have been estimated based on the following assumptions:
• PMGSY roads constructed so far: 5,29,975 km
• Ratio of premix carpet and surface dressing: 80:20
• Cost of premix carpet and surface dressing per km: ₹7.88 lakhs and ₹2.62 lakhs, respectively (based on 2012 Rajasthan PWD Basic Schedule of Rates)
• 5-year cycle of renewal of surface for existing roads.
Implementation of Surface Dressing
Individuals cannot be blamed for this collective “conventional wisdom” of using premix carpet for over 50 years despite the fact that it is highly expensive and is an undesirable technology. However, some engineers who “like” premix carpet would like to maintain the status quo and would make one or more of the following misleading statements (author’s response is in parenthesis):
1. Premix carpet is “good” for India. (It is just an unfounded vague statement; if it is good why have we not been able to convince and export this technology to rest of the world?)
2. I have seen it perform well. (Not all jobs fail, but fundamentally it is water-trapping
and therefore premature potholing is highly likely).
3. It provides structural strength to the road. (This is absolutely wrong. Both surface dressing and premix carpet do not contribute to structural strength of the road).
4. It provides a black and shiny surface and is laid sometimes with a paver, which pleases the public. (This seems to indicate public perception is more important than proper engineering and the welfare of our country in terms of costs and performance).
5. Let’s undertake some research projects to compare the performance of premix carpet versus surface dressing. (This is just a delaying tactic. Satisfactory performance of surface dressing has been proven and reported worldwide. Moreover, India’s Rural Roads Manual allows either one so why not use surface dressing which costs only one-third of premix carpet. The Manual also states that surface dressing is an age-old technique that really seals the road well).
6. Let’s set up some demonstration project to “reintroduce” surface dressing. (Again, it is just a delaying tactic. Why do it when we have an excellent and easy to follow IRC standard for surface dressing for our road engineers? Most states have surface dressing as item in their Basic Schedule of Rates. If notices inviting tenders are issued, Indian engineers/ contractors are smart to execute the contracts without any problem).
India’s Rural Roads Manual permits two technologies: surface dressing or premix carpet for bituminous surfacing of rural roads. Both have been standardized by the Indian Roads Congress. However, premix carpet is highly expensive (three times the cost of surface dressing); generally, soaks up water readily which results in early potholes; and not used anywhere in the world. Surface dressing is cheap (only onethird the cost of premix carpet); makes the road highly waterproof thus less potential for potholes; and is used successfully throughout the world. Therefore, there is no doubt that surface dressing has significant advantages over premix carpet in terms of life (performance) and cost. Besides new construction, savings of over ₹4,400 crore can be realized per year for resurfacing existing rural roads if surface dressing is used in lieu of premix carpet.
Annexure Life cycle cost analysis (LCCA) of Premix Carpet (PMC) and Bituminous Concrete (BC) Grading 2
Analysis period = 6 years
• Average life of 20 mm PMC with sand seal coat = 3 years (real 2 years)
• Average life of 40 mm BC Grading 2 = 6 years (real 7-8 years)
[This means 20 mm PMC will be required for rehabilitation of the pavement after 3 years.]
• Cost of 20 mm PMC per km lane = 7.88 lakhs
• Cost of 40 mm BC Grading 2 per km lane = 12.00 lakhs
• Real discount rate = 4%
• Net present value (NPV) = Initial cost + Rehab cost (1/(1+r)n)
• Salvage value considered equal after the 6 years analysis period (although BC would have a significantly higher structural strength whereas PMC strength is almost zero)
• Remaining service life after analysis period of 6 years considered equal (although BC would have more service life because it is dense graded)
• No maintenance activity considered during the 6-year period (although PMC is likely to require some activity)
• User operating costs considered equal (although BC would provide a smoother ride and less operating costs). Deterministic Approach was used for LCCA, which is easy and is used traditionally. The Net Present Value (NPV) was calculated for PMC and BC for the 6-year period as follows:
NPV of PMC = 14.89 lakhs
NPV of BC Grading 2 = 12.00 lakhs
This means, PMC is 24.1% more expensive than BC Grading 2. If the remaining service life, salvage value, maintenance costs, and user operating costs are considered (which all are in favour of BC), PMC would be much more expensive than 24.1 percent, which was calculated with very conservative assumptions, listed above.
About the author
Prof. Prithvi Singh Kandhal is Associate Director Emeritus of the National Center for Asphalt Technology (NCAT) based at Auburn University, Alabama, U.S.A. NCAT is the largest asphalt (bitumen) road technology center in the world. Prof. Kandhal has been a practicing highway engineer in India for over 20 years and in the US for 30 years. He has drafted many standards for the Indian Roads Congress including specifications for dense graded bituminous mixes, stone matrix asphalt and readymade pothole patching mix.
Figure 1: Schematic of surface dressing
Figure 3: Surface water entering the premix carpet (PMC) through an unsealed area saturating it under the seal as well, causing stripping within PMC and the underlying bituminous course when subjected to traffic loads
Figure 2: Typical chip spreader manufactured in Gujarat
Figure 4: Field permeability of PMC without any sand seal coat
Figure 5: Field permeability of PMC with moderate sand seal coat
Figure 6: Field permeability of PMC with adequate sand seal coat
Figure 7: Field permeability of BC Grading 2
Figure 9: Failure of premix carpet (PMC) during the first monsoon within Jaipur city
Figure 8: Steel ring used for field permeability test